Role of the Reticulospinal and Corticoreticular Systems for the Control of Reaching in
Non Human Primates.
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University
By
Lynnette Ruth Montgomery
Graduate Program in Neuroscience
The Ohio State University
2013
Dissertation Committee:
John Buford. Advisor
Lyn Jakeman
Susan Travers
D. Michele Basso
Copyrighted by
Lynnette Ruth Montgomery
2013
Abstract
The control of reaching involves an interplay of multiple motor systems within the central nervous system entailing the use of both upper limbs (UL) and the upper trunk. The purpose of this dissertation was to describe how two of these motor systems
(corticospinal and reticulospinal) contribute to the control of reaching. Firstly, we detailed the organization of reticulospinal cells within the pontomedullary reticular formation (PMRF) that project to the primate cervical enlargement in the primate cervical spinal cord. Secondly, we investigated the effects that three cortical motor areas have on both ipsilateral and contralateral motor output. Thirdly, we described the nature and organization of projections from the supplementary motor area (SMA) to the PMRF and specifically to reticulospinal cells that project to the cervical enlargement in the primate.
The reticulospinal system, which originates from the PMRF in the brainstem, has been shown to contribute to the control of reaching through its influence on muscles of the trunk and proximal UL. Not only is this system focused on more proximal musculature, but it is also a bilateral system, meaning that motor outputs can influence both UL’s. Despite this potential role in the control of skilled UL movement, there are currently few details in the literature regarding the structural organization of the reticulospinal system in the primate. Using retrograde tract-tracers injected into the cervical enlargement of non-human primates (Macaca fascicularis), we labeled
ii reticulospinal cells that projected to the cervical enlargement where alpha motoneurons for UL muscles are located. Following unilateral tracer injections, we found that reticulospinal cells from both the left and right PMRF project to the cervical enlargement, however significantly more cells were in the PMRF ipsilateral to the injection site. Most reticulospinal cells were in the nucleus reticularis pontis caudalis (PnC) and the rostral portion of the nucleus reticularis gigantocellularis (Gi); this was also where the majority of the largest cells were located.
Although there are multiple motor systems involved in the control of reaching, the corticospinal system has received the most attention due to the fact that – (1) projections originate directly from the motor cortex, (2) this system is most developed in higher order animals such as non-human primates and humans, and (3) this system is particularly important for fine motor control in primates. Because the majority of corticospinal projections are contralateral in nature, studies have focused on its role in the control of the distal contralateral UL and few studies have investigated the role that the corticospinal tract plays in the control of the ipsilateral UL, and there is also a paucity of information about corticospinal control over proximal muscles in the UL and trunk. The second study in this dissertation uses electrophysiology - intracortical microstimulation and electromyography (EMG) - to investigate the contributions that the primary motor cortex (M1), SMA, and dorsal premotor area (PMd) make to the control of the ipsilateral and contralateral UL during a reaching task. These three areas were chosen because they are the site of the majority of corticospinal cells. Stimulus trains (36 biphasic pulses, 330
Hz) were delivered to one of these three cortical areas and EMG was recorded from
iii select muscles in the upper trunk and proximal UL’s. Our findings showed that there were a number of muscle responses detected in the ipsilateral UL following SMA stimulation, especially around the trunk and shoulder girdle. Our use of stimulus trains meant that both direct and indirect motor pathways were activated. With this in mind, the fact that the onset of ipsilateral responses was delayed compared to the most direct contralateral responses following M1 stimulation indicated that ipsilateral responses were elicited through indirect pathways such as the reticulospinal pathway.
As SMA produced a number of ipsilateral responses which seemed to be elicited through indirect rather than direct motor pathways, and the reticulospinal system in the primate projects primarily ipsilaterally to the cervical spinal cord in the primate, we wanted to detail whether a pathway existed between SMA, reticulospinal cells, and the cervical enlargement. Using neuroanatomical tract-tracers (anterograde tracers in SMA and retrograde tracers in the cervical cord), we identified labeled corticoreticular boutons in the ipsilateral and contralateral PMRF on fibers that originated from the arm representation of SMA. A small but significant number of these boutons came within 5
µM of labeled reticulospinal cells, making it likely that these boutons were making contact with these reticulospinal cells. Staining with a presynaptic marker
(synaptophysin) provided further evidence that these boutons that came within 5 µM were forming functional contacts with reticulospinal cells. Although these contacting boutons were found in both sides of the PMRF, the majority of them were located in the
PMRF ipsilateral to the SMA injection. The majority of these boutons were also located in PnC and rostral Gi where most of the reticulospinal cells are located.
iv
Thus, this dissertation presents evidence that the reticulospinal system in the primate is a bilateral, although predominantly ipsilateral pathway that projects to the cervical enlargement. We also provide evidence that premotor areas of the cortex, especially SMA, elicit a number of responses in the ipsilateral UL mainly proximal muscles of the trunk and shoulder girdle, and these responses appear to be elicited through indirect polysynaptic pathways. One such pathway that may be involved in producing these effects involves projections from SMA to reticulospinal cells, which then project to the cervical cord. With our final experiment we have provided evidence that such a corticoreticulospinal pathway from SMA to the cervical cord via the reticulospinal system does exist in the primate. This pathway is bilateral, although it is predominantly ipsilateral. This corticoreticulospinal pathway has the capacity to influence both ipsilateral and contralateral UL musculature and may provide a mechanism by which UL motor recovery can take place following cortical injury.
v
Acknowledgments
I would like to thank my PhD advisor, Dr John Buford, for all of his guidance and support during this PhD experience. I would also like to offer a special thank you to Dr
Lyn Jakeman who taught me invaluable skills in the area of neuroanatomy and general science. You have allowed me to be able to take my love and passion for neuroanatomy and turn it into my career. That is a gift I can never repay. Thank you also to my other committee members, Dr Michele Basso and Dr Susan Travers, for providing constructive guidance and helping me to make it this far. A very special thank you also goes to
Stephanie Moran and Rebecca Slattery. There are many things I have learnt from you that have helped me both personally and professionally. Without you two I would not have had the confidence to get to the finish line. Finally, but most importantly, I would like to thank my family. To my husband, John, who has always been by my side and has always believed in me even when I did not believe in myself, I am eternally greatful for your support and love. To my mother, you taught me to always ask questions even when it is not the accepted thing to do and it is that drive which had driven me on to achieve the success that I have had.
vi
Vita
1993...... Graduated from Wavell Heights High
1999...... Bachelor of Physiotherapy, University of
Queensland, Australia
2000-2001 ...... Physiotherapist, Royal Hobart Hospital
2002-2003 ...... Physical Therapist, Providence Everett
Medical Center
2002-2003 ...... Per Diem Physical Therapist, Overlake
Medical Center
2003-2005 ...... Physiotherapist, Palmerston North Hospital
2005-2007 ...... Physical Therapist, St Joseph’s Hospital and
Medical Center
2007 to present ...... Graduate Research Associate, School of
Health and Rehabilitation Sciences, The
Ohio State University
vii
Publications
Montgomery LR, Herbert WJ, and Buford JA (2013) Recruitment of ipsilateral and contralateral upper limb muscles following stimulation of the cortical motor areas in the monkey. Experimental Brain Research 230(2): 153-164.
Fields of Study
Major Field: Neuroscience
viii
Table of Contents
Abstract ...... ii
Acknowledgments...... vi
Vita ...... vii
Publications ...... viii
Fields of Study ...... viii
Table of Contents ...... ix
List of Tables ...... xi
List of Figures ...... xii
Chapter 1: Introduction ...... 1
Chapter 2 : Background and Significance ...... 8
Postural control in the upper limb ...... 11
Reticulospinal tract ...... 13
Cortical Motor Areas and Corticospinal tract ...... 21
Connections between motor cortex and reticulospinal neurons ...... 26
Conclusion ...... 30 ix
Chapter 3 Anatomical organization of cells of the reticulospinal tract in the pontomedullary reticular formation of the non-human primate ...... 33
Methods ...... 36
Results ...... 46
Discussion ...... 62
Conclusion ...... 69
Chapter 4 Recruitment of ipsilateral and contralateral upper limb muscles following stimulation of the cortical motor areas in the monkey...... 71
Materials and Methods ...... 73
Results ...... 81
Discussion ...... 90
Chapter 5 : Projections from the supplementary motor area to reticulospinal cells in the
PMRF in a non-human primate as revealed by anterograde and retrograde tract tracing. 98
Materials and Methods ...... 102
Results ...... 117
Discussion ...... 135
Chapter 6 : Discussion ...... 147
References ...... 159
x
List of Tables
Table 3.1 Number of objects counted in each PMRF nucleus ...... 46
Table 3.2 Number of labeled reticulospinal cells found in each PMRF nucleus ...... 47
Table 4.1 Number of muscle responses detected...... 88
Table 5.1 Number of BDA swellings in each PMRF nucleus ...... 121
xi
List of Figures
Figure 2.1 Golgi stains of reticulospinal cells ...... 14
Figure 2.2 Muscle responses following PMRF stimulation ...... 16
Figure 2.3 The classical homunculus map...... 22
Figure 2.4 Corticoreticular projections ...... 28
Figure 3.1 Retrograde tracer injections sites...... 40
Figure 3.2 Distribution of labeled reticulospinal cells in each of the four PMRF nuclei 48
Figure 3.3 Proportion of reticulospinal cells in each nucleus ...... 49
Figure 3.4 Total estimated number of cells found ipsilaterally and contralaterally ...... 50
Figure 3.5 Distribution of reticulospinal cells throughout each PMRF nucleus for subject
A...... 52
Figure 3.6 Distribution of reticulospinal cells throughout each PMRF nucleus for subject
G...... 53
Figure 3.7 Distribution of reticulospinal cells throughout each PMRF nucleus for subject
R...... 54
Figure 3.8 Photomicrographs of small, medium, large and giant cells...... 55
Figure 3.9 Distribution of the large and giant reticulospinal cells throughout each PMRF nucleus for subject A...... 58
xii
Figure 3.10 Distribution of large and giant reticulospinal cells throughout each PMRF nucleus for subject G...... 59
Figure 3.11 Distribution of the large and giant reticulospinal cells throughout each PMRF nucleus for subject R...... 60
Figure 4.1 Detecting suppression events...... 79
Figure 4.2 Responsive stimulation sites in each cortical motor area...... 82
Figure 4.3 EMG response from a stimulus train applied in SMA...... 83
Figure 4.4 Muscle responses by laterality for each of the cortical motor areas...... 85
Figure 4.5 Ipsilateral and contralateral responses based on UL region...... 89
Figure 5.1 Anterograde injection sites in the left and right SMA...... 107
Figure 5.2 Schematic representation of how the ROI contour for each section for each
PMRF nucleus was constructed...... 115
Figure 5.3 BDA swellings in each PMRF nucleus for each subject...... 119
Figure 5.4 Distribution of BDA swellings and reticulospinal cells throughout the four
PMRF nuclei for subject A...... 125
Figure 5.5 Distribution of BDA swellings and reticulospinal cells throughout the four
PMRF nuclei for subject G...... 126
Figure 5.6 Location of BDA swellings and location of reticulospinal cells in the ipsilateral PMRF...... 128
Figure 5.7 Probable contacts...... 130
Figure 5.8 Distance of swellings from the cell plotted on the graph ...... 131
xiii
Figure 5.9 Number of cells that received probable contacts from corticoreticular swellings...... 132
Figure 5.10 Relationship between cell size and probable contacts …………………134
Figure 5.11 Schematic diagram of the corticoreticulospinal pathway…………………136
xiv
Chapter 1: Introduction
The role of the central nervous system in controlling the upper limb has been the focus of much interest, especially in primates where the ability to perform fine dexterous hand movements has received particular attention (Lemon, 2008b; Maier and others,
1993; Schieber and Santello, 2004). Due to this emphasis on hand function, most studies involving motor control of the upper limb are focused on the corticospinal tract, in particular the primary motor cortex (M1), which is the source of most direct projections from the cortex to alpha motoneurons supplying muscles involved in hand and finger movement (Boudrias and others, 2010b; Cheney and Fetz, 1985; Fetz and Cheney, 1980;
Fetz and others, 1976; Lemon and Griffiths, 2005; Lemon and others, 2004). However, hand and finger movements can only be successfully performed if the rest of the arm provides a stable platform (Castiello and others, 1992; Gentilucci and others, 1992;
Huang and Brown, 2013). This function of limb stability has received limited attention in the literature to date. Studies have focused on outputs from the supplementary motor area (SMA), a secondary motor area in the cortex, which is associated with activity in proximal muscles and those muscle groups which are involved in postural control
(Hummelsheim and others, 1986; Jacobs and others, 2009; Wiesendanger and others,
1974). In addition to motor pathways from the cortex, alternate motor pathways from
1 subcortical areas such as the pontomedullary reticular formation (PMRF) have also been shown to play a role in proximal upper limb activity and postural control (Buford and
Davidson, 2004; Davidson and Buford, 2006; Prentice and Drew, 2001; Schepens and
Drew, 2003; Schepens and Drew, 2004). Though there is evidence that these pathways from these two areas may play a similar role in upper limb stability, no one has yet investigated this connection within the primate model. This dissertation represents a body of work that investigates how motor control of the upper limb may be influenced by the reticulospinal tract and SMA in the non-human primate (NHP). Firstly the anatomical organization of the reticulospinal system within the NHP is described in chapter 3. In chapter 4, the role of SMA and other cortical motor areas in the recruitment of proximal muscles in the ipsilateral and contralateral upper limbs. Finally in chapter 5, evidence is presented detailing the anatomical organization of the corticoreticular pathway from SMA to reticulospinal cells within the PMRF.
Previous studies from our lab have shown that cells in the PMRF can elicit activity in muscles of the trunk and proximal upper limb during reaching movements, suggesting that the reticulospinal system preferentially controls these more proximal, postural muscle groups that are associated with the postural control of reaching (Buford and Davidson, 2004; Davidson and Buford, 2006). Additionally these outputs have been shown to be bilateral, affecting both upper limbs. In the case of a neurological injury, this provides a motor pathway through which the upper limb muscles in the paralyzed arm could be activated following a unilateral injury to the motor cortex or corticospinal tract. The question is, how is the reticulospinal pathway organized so that it can elicit
2 these bilateral responses? Although the anatomical organization of the reticulospinal tract has been detailed for many animal models, showing that the reticulospinal tract is bilateral although predominantly ipsilateral with its projections to the spinal cord (Beran and Martin, 1971; Jankowska and others, 2003; Matsuyama and others, 1999; Matsuyama and others, 1997; Newman, 1985a; Newman, 1985b; Peterson, 1979; Peterson and others,
1975; Peterson and others, 1979), the organization of the reticulospinal tract in the non- human primate remains understudied. With the substantial changes in the control of the upper limb that is evident in higher primates, some changes in the organization of the reticulospinal tract in primates might be expected. Is the tract mostly ipsilateral as it descends in the spinal cord producing bilateral effects via commissural spinal interneurons as in other animal models (Bannatyne and others, 2003; Jankowska and others, 2003)? Or, does the reticulospinal tract in the primate project bilaterally in the spinal cord, thus requiring minimal input from commissural spinal interneurons? By understanding the anatomical organization of the reticulospinal tract in the intact primate nervous system, science will be able to more clearly identify where plasticity such as axonal sprouting may occur within the system following injury. If the bilateral effects occur mainly through commissural spinal neurons, then sprouting is most likely to occur in the spinal cord. If the bilateral effects occur following reticulospinal branching in the brainstem, then most likely sprouting would occur in the brainstem. An earlier study from our lab where we attempted to detail the organization of the reticulospinal tract
(Sakai and others, 2009) led to some unexpected findings that indicated that reticulospinal projections were relatively bilateral in the spinal cord. As this finding is in
3 contrast to the organization of the reticulospinal tract in other animal models, we wanted to confirm these findings using unbiased stereological techniques. Chapter 3 of this dissertation uses tract tracing and stereological techniques to demonstrate that the organization of the reticulospinal tract is conserved in the non-human primate, with most of the projections being ipsilateral to the spinal cord. A majority of the reticulospinal cells originate nucleus pontis caudalis (PnC) and the rostral part of nucleus gigantocellularis (rGi) within the PMRF. These two nuclei are where the majority of large and giant reticulospinal cells are located, which have been shown to receive input from multiple centers of the brain including motor and sensory systems and the cerebellum (Eccles and others, 1975; Matsuyama and Drew, 1997; Scheibel and others,
1955). These larger cells contribute to an important postural response called the acoustic startle reflex (Lingenhohl and Friauf, 1992; Yeomans and Frankland, 1995). Thus it appears that the organization of the reticulospinal tract in primates is primed for the function of proximal stability and postural control during movement as it is in lower order animals.
Many in the area of rehabilitation research have examined the role that the cortical motor areas may have in the recovery of function following cortical injury, focusing on the potential for ipsilateral projections from the motor cortex to contribute to the recovery of function (Bradnam and others, 2013; Rosenzweig and others, 2009;
Rosenzweig and others, 2010). There is, however, a paucity of literature detailing the functional contributions that three of the prominent cortical motor areas make to both the ipsilateral and contralateral upper limb in the non-human primate, especially with regards
4 to the proximal musculature. Chapter 4 of this dissertation addresses this paucity in the literature by detailing the pattern of muscle activity in the proximal upper limb following stimulation of M1, SMA, and the dorsal premotor area (PMd). Stimulation of all three areas produced contralateral and bilateral muscle activity in the upper limb as expected.
Interestingly though, stimulation of the premotor areas (SMA and PMd) resulted in some instances of purely ipsilateral muscle activity. This ipsilateral activity was most common in the most proximal muscle groups around the upper trunk and shoulder girdle. There were also significantly more ipsilateral responses following stimulation of SMA. The onset of the responses from PMd and SMA were significantly later than responses from
M1, indicating that the pathways from PMd and SMA were more likely to be multisynaptic, involving neurons in the brainstem or spinal cord before reaching the motoneurons. Thus SMA and PMd may provide a mechanism by which the motor cortex may be involved in postural control of both arms during upper limb movements, given their preference to activate bilateral muscle groups in the trunk and shoulder girdle.
Given that these responses seem to be delayed compared with more direct activation from
M1, the idea that SMA and PMd responses may in part be facilitated through alternate motor pathways in the brainstem, such as the reticulospinal system, is plausible.
As SMA has been shown to contribute significantly to postural responses that occur prior to movement and in response to movement (Hummelsheim and others, 1986;
Tanji and others, 1988; Wiesendanger and others, 1974), a function similar to the reticulospinal tract, the possibility that a pathway from SMA to the PMRF and then to the spinal cord may exist (a corticoreticulospinal pathway) seems natural. Although there is
5 some evidence that such a pathway exists (Keizer and Kuypers, 1984; Keizer and
Kuypers, 1989; Matsuyama and Drew, 1997; Newman and others, 1989; Rho and others,
1997), there has yet to be a detailed investigation of such a pathway in the non-human primate. Chapter 5 of this dissertation presents evidence of the existence of corticoreticular projections from SMA to the PMRF and describes the organization of this pathway in the non-human primate using anterograde and retrograde tract tracers. SMA sends bilateral projections to the PMRF and these projections are concentrated in the medial and intermediate portions of the PMRF nuclei where a majority of the reticulospinal cells are found. Of these corticoreticular projections, there are a small but significant number of SMA projections that come close enough to reticulospinal cells to make functional connections. These probable contacts were predominantly found in the ipsilateral PMRF. Larger reticulospinal cells were more likely to receive more probable contacts than smaller cells. As noted earlier, the larger reticulospinal cells play a vital role in the postural control component of the reticulospinal tract, and given that SMA is also important in proximal muscle activity and postural control, these data are consistent with the hypothesis that the corticoreticulospinal pathway from SMA to the spinal cord via the PMRF is part of a system that provides proximal stability during upper limb movements such as reaching.
This dissertation provides evidence that projections from one cortical motor area
(SMA) and the reticulospinal tract can cooperate to control upper limb movement.
Indications are that this function is primarily postural in nature involving the control of proximal musculature to provide a stable platform on which more distal movement can
6 occur. Many take this function for granted due to the fact that such movement is often subtle and not readily observable. However, following a neurological injury, the recovery of hand function (which is often a primary goal of rehabilitation) is often incomplete due to the inability to have this stable proximal platform. The importance of this proximal stability is also often overlooked in rehabilitation, leading to ongoing disability and decreased quality of life. Following neurological injury, a lack of stability around the shoulder often results in soft tissue injuries such as shoulder subluxation which can lead to intractable pain (Ada and others, 2005; Lindgren and others, 2007;
Pertoldi and Di, 2005). Could SMA and the reticulospinal system play an integral part in this postural control component of reaching, and if so, how can we use this system to facilitate more complete recovery of upper limb function following cortical injury? This fundamental clinical question can best be addressed with a more thorough understanding of how these two systems may influence postural control during reaching in a non-injured system, including how these systems may interact to facilitate postural stability of the shoulder during movement. The studies included in this dissertation start to address these gaps that are currently in the literature. More research into the interactions between the cortical motor areas and reticulospinal tract is needed in order to elucidate the mechanisms by which these two systems may contribute to proximal stability of the upper limb and how rehabilitative strategies may be used to maximize this function following cortical injury such as stroke.
7
Chapter 2 : Background and Significance
In human beings, the ability to move the arm and hand to reach and grasp for objects requires the coordination of a number of motor pathways within the nervous system (Baker and others, 2001; Brochier and others, 2004; Lemon, 2008b; Lemon and others, 2004; Spinks and others, 2008). Upper limb function is a vital component of many activities of daily living such as getting dressed, grooming and self-care, eating, and drinking. In particular, the ability to grasp and manipulate objects is seen as the most important part of these tasks and as a result recovery of hand function is often the centerpiece of upper limb rehabilitation programs following neurological injury.
However, the proximal arm, shoulder girdle, and trunk are also active during upper limb tasks, in order to provide a stable platform from which hand movement can occur
(Castiello and others, 1992; Gentilucci and others, 1992; Huang and Brown, 2013).
The corticospinal tract is crucial for skilled reaching, especially with regards to the performance of complex movements of the hand (Asanuma, 1973; Boudrias and others, 2010b; Cheney and Fetz, 1985; Fetz and Cheney, 1980; He and others, 1993;
Kwan and others, 1978; Maier and others, 1993; Ralston and Ralston, III, 1985; Schieber,
1990; Schieber, 2001). Although the corticospinal tract does contribute to the proximal control of the upper limb as well, alternate motor pathways such as the reticulospinal
8 tract also play an important role in the control of the proximal upper limb which allows distal hand movement to be successful (Davidson and Buford, 2006; Drew and
Rossignol, 1990a; Schepens and Drew, 2004). Owing to the ability of the corticospinal tract to control intricate, fine, dexterous hand movements and the importance of hand function in activities of daily living, this tract has received the most attention in the literature. In contrast, the importance of proximal stability in activities of daily living and the contributions of the alternate motor pathways such as the reticulospinal tract to upper limb function has been poorly studied to date. This paucity in the literature has led to limited approaches to rehabilitation of the upper limb following neurological injury, with an emphasis on recovery of distal upper limb function without adequate attention to proximal upper limb recovery (Donaldson and others, 2009; Kwakkel and others, 1999).
This in turn results in incomplete recovery of upper limb function and an inability to perform many activities of daily living effectively and efficiently (Broeks and others,
1999). It is therefore imperative that there be a greater understanding of the role that the corticospinal and reticulospinal tracts play in the proximal stability (postural control) of the upper limb, so that improved rehabilitation strategies may be developed that will lead to a more comprehensive recovery of upper limb function.
Most of the studies which have investigated the reticulospinal tract have been undertaken in lower order animals such as cats, rodents, and opossums (Beran and
Martin, 1971; Canedo and Lamas, 1993; Drew and others, 1986; Martin and Dom, 1971;
Martin and others, 1979; Matsuyama and others, 1993; Matsuyama and others, 1999;
Matsuyama and others, 1997; Peterson, 1979; Peterson and others, 1979; Peterson and
9 others, 1978; Schepens and Drew, 2004; Schepens and Drew, 2006; Stapley and Drew,
2009). Although this has provided important insights into the overall structure and function of the reticulospinal system across various species, in order to address questions about the role of the reticulospinal system in the control of reaching, a different animal model is required. The upper limb is primarily a weight bearing structure in lower order animals, unlike its function in primates where it is involved not only in locomotion, but also grasping and manipulating objects (Heffner and Masterton, 1983; Lemon and
Griffiths, 2005; Quallo and others, 2012). This means that our ability to translate findings from studies involving lower order animal models to human upper limb function may be limited. The most outstanding feature that separates upper limb function of primates from lower order animals is the ability to perform more precise fine movements of the hand and fingers due to the presence of direct projections from the motor cortex to alpha motoneurons serving the hand (Bortoff and Strick, 1993; Fetz and Cheney, 1980; Lemon and others, 2004; Nakajima and others, 2000; Rathelot and Strick, 2006). Indeed, most studies detailing corticospinal control of upper limb function use non-human primate models due to the similarities in upper limb function and cortical representation between non-human primates and humans (Lemon and Griffiths, 2005; Lemon and others, 2004;
Quallo and others, 2012). Thus, any attempt to understand how the reticulospinal and corticospinal tracts may interact to control the upper limb is best suited to a non-human primate model due to the increased complexity of corticospinal projections in primates and similarities in upper limb function.
10
Postural control in the upper limb
For many, the term postural control involves the ability to maintain an upright position against gravity (i.e. the ability to not fall over). Such a task involves activation of muscles in the trunk and weight bearing limbs like the legs. Postural control, however, is a broader term that not only includes the well-known definition described above, but also the ability to maintain a limb posture during movement (Flash and Mussa-Ivaldi,
1990; Massion, 1994). All movements expend energy, and the ability to control limb posture during movement can be seen as a way in which the nervous system is able to expend the least energy possible while performing a movement (Soechting and others,
1995). This postural control is often overlooked during movement because the muscle activity required is subtle and often unobservable. Following a neurological injury, the ability to move without expending excessive amounts of energy is often lost (Houdijk and others, 2010). This is one of the factors that can lead to long term disability and decreased quality of life for these injured individuals. Additionally, the lack of postural control around the shoulder following neurological injury often leads to soft tissue injuries such as shoulder subluxation, and in some extreme case shoulder-hand syndrome, which can lead to intractable pain (Ada and others, 2005; Lindgren and others, 2007;
Pertoldi and Di, 2005). Even if these patients recover some movement around the shoulder, it is typically gross, poorly regulated movement with abnormal muscle tone, which may be one factor leading to the aforementioned problems of soft tissue injuries and persistent pain (Dewald and others, 1995). Despite these impairments, few have
11 researched the neural substrates involved in the control of limb posture, and thus our rehabilitation approaches to address these problems are very limited.
The shoulder complex is a relatively unstable structure, designed for mobility at the expense of stability. Most of the stability for the shoulder originates from muscles surrounding it, in particular a group of muscles called the rotator cuff (Labriola and others, 2005). Thus, the proper recruitment of the muscles around the shoulder during arm and hand tasks is vital for success. As mentioned previously, this muscle activity often occurs in the background, seemingly unrelated to the actual task that is being performed. As a result, the mechanisms involved in performing this proximal stabilization part of the task are not well-known.
One area in the nervous system that seems particularly important for the recruitment of proximal muscles around the shoulder and upper trunk is the supplementary motor area (SMA) (Kurata and Tanji, 1986; Macpherson and others,
1982; Tanji and Kurata, 1979). Stimulus trains applied to SMA result in a number of muscle responses in the upper trunk and shoulder girdle (Montgomery and others, 2013).
Not only were a number of responses seen in these proximal muscles, but the responses were bilateral, occurring in muscles in the ipsilateral and contralateral upper limbs
(Montgomery and others, 2013). These characteristics of SMA activity show striking similarities to muscle responses related to reticulospinal activity. Physiological studies from our lab (Davidson and Buford, 2006; Herbert and others, 2010) have also demonstrated that reticulospinal inputs are pronounced to proximal muscles around the shoulder and upper trunk, especially to muscles of the rotator cuff such as supraspinatus.
12
These responses are also bilateral, with responses seen in both the ipsilateral and contralateral arms. Thus these two pathways – corticospinal tract projections from SMA, and the reticulospinal tract –play a role in activating bilateral proximal muscles around the shoulder girdle and upper trunk that involved in the postural control of the upper limb during arm and hand movements. However, even though SMA and the reticulospinal tract have similar roles in the motor control of the upper limb, few have looked at how these two structures may interact to control the upper limb. Anatomical findings imply that some connection exists between SMA and the reticulospinal tract (Keizer and
Kuypers, 1989), however the details about how these systems may interact remains unknown.
Reticulospinal tract
Pontomedullary reticular formation and reticulospinal tract
The reticulospinal tract is comprised of projections of axons from cells in the pontomedullary reticular formation (PMRF) in the pons and medulla (Brodal, 1956;
Newman, 1985a; Newman, 1985b; Valverde, 1961). These axons project to all levels of the spinal cord and terminate in the ventral gray matter, mainly on interneurons in lamina
VII and VIII, although some fibers do terminate on motoneurons in lamina IX
(Jankowska and others, 2003; Martin and Dom, 1971; Martin and others, 1979;
Matsuyama and others, 1993; Matsuyama and others, 1999; Matsuyama and others, 1997;
Peterson, 1979; Peterson and others, 1975). There are three nuclei within the PMRF
13 where reticulospinal cells are located – nucleus reticularis pontis oralis (PnO), nucleus reticularis pontis caudalis (PnC), and nucleus reticularis gigantocellularis (Gi). The rostral portion of Gi (rGi) and PnC are the primary regions where most of the reticulospinal activity related to upper limb function is located (Buford and Davidson,
2004; Davidson and Buford, 2006; Drew and Rossignol, 1990a; Drew and Rossignol,
1990b; Herbert and others, 2010). The morphology of cells in the reticular formation has been studied in detail using Golgi’s impregnation (Brodal, 1956; Scheibel and others,
1973; Valverde, 1961). One feature of reticulospinal cells is their complex dendritic arborization and the amount of territory which these dendrites span, shown in figure 2.1.
Figure 2.1 Golgi stains of reticulospinal cells show the morphological characteristics of these cells in the cat (reproduced from Scheibel and others 1973). Of particular note is the extensive dendritic arborization present on most reticulospinal cells as well as the networks that they form with other neighboring reticulospinal cells as shown in this figure. The drawing was made at 160 x.
14
Reticulospinal dendrites can extend up to 800µM in length and form complex networks with other nearby reticulospinal cells (Scheibel and others, 1973; Valverde, 1961). As reticulospinal cells receive multiple inputs from a number of other neural systems including sensory systems and the cerebellum, this complex dendritic organization provides a larger surface area on which these inputs can be made (Scheibel and others,
1955). By receiving inputs from multiple systems, reticulospinal cells appear to have an integrative function within the nervous system, allowing them to take information from many different systems and produce a single motor response. This is an important feature of reticulospinal cells which makes them adept at facilitating postural responses associated with limb movements.
A primary function of the reticulospinal tract involves assisting with locomotion and postural control through eliciting muscle responses in the trunk and limbs (Drew and others, 1986; Grillner and others, 1995; Mori and others, 1995; Mori and others, 1998;
Prentice and Drew, 2001). These functions have led many researchers to focus on the effect of reticulospinal activity on the trunk and lower limbs, as the upper limb was considered to have no involvement in postural control. Although this view has been changing and more research is now being done on the effects of reticulospinal activity on upper limb musculature (Baker, 2011; Buford and Davidson, 2004; Davidson and Buford,
2006; Schepens and Drew, 2003; Schepens and Drew, 2004; Schepens and Drew, 2006), much is still unknown about the role of the reticulospinal tract in the control of reaching in the intact and injured nervous system.
15
Reticulospinal tract and muscle recruitment
Physiological studies in various animal models have shown that neural activity in reticulospinal cells results in bilateral changes in muscle activity around the upper trunk and proximal arm (Buford and Davidson, 2004; Davidson and Buford, 2006; Peterson and others, 1979; Schepens and Drew, 2006). Further analysis demonstrates that there is an underlying stereotypical pattern of muscle activation associated with reticulospinal
Figure 2.2 Electrical stimulation of cells in the PMRF often produce a distinct pattern of muscle activity in the upper limbs (reproduced from Davidson and Buford, 2006). Each point for each muscle represents the percentage of post stimulus effects that were facilitation (e.g. the higher the percentage the more facilitation post stimulus effects were seen in that muscle). The graph shows that in the muscles ipsilateral to the PMRF stimulation (open triangle) flexor muscles demonstrated a higher percentage of facilitation than extensors. The opposite was true for muscles contralateral to the stimulation (dark squares), where extensor muscles had a higher percentage of facilitation than flexors. The abbreviations for the muscles are as follows - FCR – flexor carpi radialis, Brac – brachialis, Bic – biceps brachii, ADlt – anterior deltoid, PMj – pectoralis major, ECU – extensor carpi ulnaris, TrLa – lateral head of triceps brachii, TrLo – long head of triceps brachii, PDlt – posterior deltoid, Lat – latissimus dorsi, MTr – middle trapezius, and UTr – upper trapezius.
16
activity, with facilitation of flexors and suppression of extensors ipsilateral to the reticulospinal activity and facilitation of extensors and suppression of flexors contralaterally (Buford and Davidson, 2004; Drew and Rossignol, 1990a; Drew and
Rossignol, 1990b). This pattern of muscle recruitment following stimulation of the
PMRF is illustrated in figure 2.2. Elements of this complex, bilateral pattern of action result from stimulation at individual sites throughout the PMRF; there is no single site that recruits flexors and another for extensors. From any given site, widespread bilateral effects are the norm for PMRF motor output to the upper limb.
Although there is extensive physiological evidence that the reticulospinal tract produces bilateral motor outputs, the data from neuroanatomical studies does not completely explain how this occurs. Studies in rats, opossums, and cats have shown that reticulospinal projections to the spinal cord are mainly ipsilateral with few bilateral projections (Beran and Martin, 1971; Martin and Dom, 1971; Martin and others, 1979;
Matsuyama and others, 1993; Matsuyama and others, 1999; Newman, 1985a; Newman,
1985b). In non-human primates, only two studies have studied the projection pattern of the reticulospinal tract and the conclusions from these studies are contradictory (Kneisley and others, 1978; Sakai and others, 2009). Kneisley and colleagues found that reticulospinal projections were mostly ipsilateral from the PMRF to the spinal cord with few contralateral projections (Kneisley and others, 1978). This is consistent with the findings from other animal models as outlined previously, suggesting that the organization of the reticulospinal tract is conserved across species. However, Kneisley’s
17 work only provided a qualitative descriptive account of the pattern of reticulospinal cell distribution. In our lab, Sakai led a study using modern tract tracing techniques to label reticulospinal cells in the ipsilateral and contralateral PMRF following unilateral tracer injections in the cervical spinal cord of non-human primates (Sakai and others, 2009). In contrast to Kneisley, Sakai’s work found that the projections were mostly bilateral with similar numbers of reticulospinal cells sending projections in the ipsilateral and contralateral PMRF. Whether this difference in results reflected methodological variations between the studies, and which results were reproducible required further study. Thus the organization of the reticulospinal tract in non-human primates requires further exploration to definitively determine whether it follows a conserved pattern of organization that is seen in other animal models, or if there is a difference in the organization of the reticulospinal tract in primates.
Based on neuroanatomical findings from lower order animal models suggesting that reticulospinal projections are mostly ipsilateral, the question becomes how a predominantly ipsilateral tract produces the relatively balanced bilateral effects that are seen physiologically. There are currently two hypothesized mechanisms by which this could occur – (1) through commissural interneurons in the spinal cord, and (2) through interneurons within the PMRF (Bannatyne and others, 2003; Jankowska and others,
2003; Peterson and others, 1975). The prevailing opinion in the literature at this time is that the pathway from the PMRF to the alpha motoneurons involves spinal interneurons, some of which are commissural, at the spinal level where the motoneuron is located
(Bannatyne and others, 2003; Jankowska and others, 2003). Jankowska and colleagues
18
(Bannatyne and others, 2003; Jankowska and others, 2003) demonstrated that reticulospinal projections often relay through commissural interneurons in the spinal cord, indicating that even if reticulospinal projections are predominantly ipsilateral, the connections they form at the spinal segmental level could produce bilateral effects.
Thus the reticulospinal tract is a motor pathway that elicits muscle responses bilaterally in the upper limb, especially in muscles around the shoulder girdle and trunk.
This bilateral action is facilitated by spinal interneurons in various animal models. The precise neuroanatomical organization of the reticulospinal tract is currently not clear, although physiologically the bilateral muscle responses that result from reticulospinal activity in lower order animals are also present in primates.
Reticulospinal tract and postural control of the upper limb
The acoustic startle reflex is a postural response that relies on reticulospinal activity to activate postural muscles (Koch and Schnitzler, 1997; Lingenhohl and Friauf,
1992; Rothwell, 2006; Yeomans and Frankland, 1995; Yeomans and others, 2002). This reflex involves activation of select muscles of the trunk and limbs in response to an unanticipated loud sudden noise. Lingenhohl and Friauf were able to show that the largest reticulospinal cells in PnC are the ones responsible for eliciting the postural responses associated with this reflex (Lingenhohl and Friauf, 1992). Another motor postural response which is associated with reticulospinal activity is the asymmetric tonic neck reflex (ATNR), a reflex that is thought to involve both the vestibulospinal and reticulospinal tracts (Denny-Brown, 1966). This reflex is initiated by turning the head to
19 one side which then activates the vestibular apparatus in both ears, sending neural signals to the vestibular nuclear complex and the reticular formation. The result is activation of extensors in the arm and leg ipsilateral to the head rotation, and activation of flexors in the arm and leg contralaterally. This pattern of arm activity resembles the pattern of muscle activation seen following PMRF stimulation. These postural responses are often used as ways to facilitate muscle tone and help patients make initial attempts at voluntary movement during early recovery from hemiparesis (Ellis and others, 2012; Lee and others, 2009).
The acoustic startle reflex and ATNR are subcortical reflexes elicited in response to a specific stimulus. In cats, it has also been shown that there is some PMRF cell activity prior to the animal performing a reaching movement (Schepens and Drew, 2003;
Schepens and Drew, 2004; Schepens and Drew, 2006). This activity is associated with changes in the loading of the limb involved with reaching, with the reaching limb unexpectedly bearing more weight instead of being unloaded (Schepens and Drew,
2003). This activity appears to be an anticipatory postural adjustment (APA), where muscles are activated in anticipation of the upcoming movement and the perturbation that is associated with this movement. These APAs, referred to by Schepens and Drew as anticipatory postural adjustments preceding the movement (pAPAs), are instrumental in providing postural stability during movement such as reaching. However, often the details of the movement change during execution, based on internal and external factors
(Bouisset and Zattara, 1987; Cordo and Nashner, 1982; Marsden and others, 1981). To cope with these unanticipated changes to the movement, dynamic postural adjustments
20
(or APAs accompanying movement as referred to by Schepens and Drew) must be made during the movement to maintain postural stability and allow for the successful completion of the task (Marsden and others, 1981; Massion, 1992; Massion, 1994). The cells that were recorded from for these postural studies were large, with average conduction velocities around 90 m/s, and they were located in the two nuclei where most of the large and giant reticulospinal cells are located – PnC and rGi (Lingenhohl and
Friauf, 1992; Newman, 1985a; Newman, 1985b; Sakai and others, 2009; Schepens and
Drew, 2004). The muscles that had the highest correlation with this postural adjustment activity were in the proximal arm around the shoulder girdle (Schepens and Drew, 2004).
Though the authors saw both APAs and dynamic postural adjustments, the greatest number of neurons studied showed preferred activity during the reach, indicating that these were either participating in the movement itself or related to dynamic postural adjustments (Schepens and Drew, 2004). Given that most of the cells were large and previous studies have shown that the largest reticulospinal cells are the ones involved with eliciting other postural responses, the hypothesis that these cells were contributing to dynamic postural adjustments seems the most plausible.
Cortical Motor Areas and Corticospinal tract
M1
The corticospinal tract consists of projections from a number of areas of the frontal lobe (cortical motor areas) as well as a small number of projections from the
21 sensory cortex in the parietal lobe (Dum and Strick, 1991; He and others, 1993; He and others, 1995). Three areas that provide a majority of the corticospinal projections are the primary motor cortex (M1), SMA, and the premotor area (PMA). Of these, M1 has been researched heavily owing to its high level of development in primates, especially humans, and its role in the control of fine dexterous movements of the hand (Lemon, 2008b; Maier and others, 1993). Work from Sherrington in the early 20th century, and Woolsey and
Penfeld in the middle part of the 20th century led the classical homunculus map of M1 in primates and humans (Lemon, 2008a; Penfield, 1954; Welker and others, 1957; Woolsey and CHANG, 1947). Figure 2.3 shows an illustration of the homunculus map. Using
Figure 2.3 The classical homunculus map created by Sherrington following stimulation of a gorilla cortex (reproduced from Lemon 2008a). This schematic diagram shows the representation of body parts in an organized pattern within M1. Stimulation of the medial areas of M1 results in movement of the lower limb. As the stimulation moves further laterally in M1 lower limb movement is replaced by trunk movement and then upper limb movement as the stimulation is applied to sites even more lateral. Stimulation of the most lateral sites in M1 elicit facial movements. A is a depiction of the sites mapped with the numbers representing the movements that resulted from faradic stimulation applied to the site, and B is a simplified map where movements were grouped according to body part.
22 electrical stimulation, researchers were able to show that different areas within M1 are associated with movement in different parts of the body. Certain body parts were also noted to have a greater representation than others, with more cortical tissue being dedicated to movement of the hand and fingers than other larger body parts such as the trunk and lower limb. Another feature of this map was that the movement elicited occurred contralaterally. Decades later, following the development of more specific stimulation techniques such as intracortical microstimulation (ICMS), it was discovered that certain parts of M1 also responded to lower currents than other parts (Asanuma and
Rosen, 1972; Jankowska and others, 1975; Kwan and others, 1978). Coincidentally the areas that responded to the lowest thresholds were associated with hand and finger movement where currents as low as 5µA could elicit visible movement (Asanuma and
Rosen, 1972). Studies have also found that there are specific cells (corticomotoneuronal cells) within M1 that project directly from M1 to alpha motoneurons (Fetz and Cheney,
1980; Fetz and others, 1976; Lemon and others, 2004; Rathelot and Strick, 2006;
Schieber, 1990; Schieber, 2001).
PMA
The premotor areas are divided into two separate areas – the ventral premotor area
(PMv) and the dorsal premotor area (PMd). These areas are secondary association areas involved with preplanning movements in particular visually guided movements such as catching a ball (Mushiake and others, 1991). Stimulation of PMd does elicit some muscle responses but the currents required to produce movement are much higher than
23 those that elicit movement in M1 (Weinrich and Wise, 1982). Movements from PMA tend to be related to more distal arm function involving grasping and manipulating objects, although there are some cells that are responsive during more proximal arm movements (Kraskov and others, 2011; Rizzolatti and others, 1988). Studies of PMA have focused on its role in movement preparation as this is the primary function of this cortical area. Cell activity is increased in PMA during the instruction and planning period of an instructed delay reach task as well as during the visual appearance of the cues and the targets (Kurata and Tanji, 1986; Tanji, 1985; Tanji and others, 1996).
Recently, research has shown that PMv contains mirror neurons which show an increased activity while an animal is performing a task as well as when that animal is watching someone else perform the same task (Kraskov and others, 2009; Rizzolatti and others,
1996). Thus, activity in PMA neurons are driven mostly by visual cues to prepare for an upcoming movement involving mostly distal grasping and manipulation of objects.
SMA
SMA was also first described in detail by Woolsey and as its name suggests it is considered a supplemental motor area in deference to M1 (Woolsey and others, 1952).
Stimulation of SMA produces observable movement, however as with PMA larger currents are required to produce these movements compared with M1 (Weinrich and
Wise, 1982). In addition to increased stimulus current being needed elicit movement, the length of stimulation must also be longer in order to elicit observable movement (Mitz and Wise, 1987). There is also evidence that a loosely organized movement map exists
24 in SMA with leg responses found most caudally and face responses located most rostrally, with arm responses in between these two (Luppino and others, 1991; Mitz and
Wise, 1987). Another difference between M1 and SMA is that whereas M1 is involved heavily in the production of muscle activity in the hand, SMA seems to show a preference for activation of more proximal muscles around the shoulder (Hummelsheim and others, 1986). Compared to M1, SMA has also been shown to produce some activity associated with ipsilateral movement (Brinkman and Porter, 1979; Hoshi and Tanji,
2004; Tanji and others, 1988). There are also numerous connections between SMAs in each hemisphere through transcallosal projections, and these connections are thought play an important role in the recovery of function following stroke (Kimberley and others, 2006; Mintzopoulos and others, 2009).
Although SMA stimulation produces movement, it is the role of SMA in the preparation of movement that has been researched most thoroughly. SMA is important in producing internally guided movements (movements that are performed automatically without the need for visual cues) and has strong connections with the basal ganglia
(Mushiake and others, 1991). Studies investigating neural activity in SMA during instructed delayed reaching tasks have demonstrated that a number of cells in SMA are involved in producing postural responses such as APAs. Cells will show an increase in activity during the time between when the decision to reach for a target is made and the actual reaching movement occurs (Hoshi and Tanji, 2004; Tanji and Shima, 1994; Tanji and others, 1996). This increased neural activity is often not associated with observable movement, however electromyographic (EMG) recordings from muscles within the trunk
25 and arm show that some muscle activity is involved (Brinkman and Porter, 1979; Mitz and Wise, 1987; Tanji and others, 1988). Wiesendanger and colleagues found that stimulation of SMA resulted in a slow contraction of proximal muscles around the shoulder in complex patterns which they concluded was consistent with postural responses (Wiesendanger and others, 1974). In humans, cortical activity associated with movement preparation (called the Bereitschaftspotential) has been shown over SMA
(Deecke, 1987; Shibasaki and Hallett, 2006; Yazawa and others, 1998). Therefore just like the reticulospinal tract, SMA seems to play a significant role in the production of
APAs and is involved with activation of predominantly proximal muscles around the shoulder.
Connections between motor cortex and reticulospinal neurons
Reticulospinal cells are integrating units, receiving inputs from various areas of the central nervous system which the cells then use to produce one motor output. In addition to the sensory and cerebellar inputs that reticulospinal cells receive, these cells also receive substantial input from the motor cortex (Kably and Drew, 1998; Keizer and
Kuypers, 1984; Keizer and Kuypers, 1989; Matsuyama and Drew, 1997; Newman and others, 1989; Rho and others, 1997). These projections from the motor cortex to the
PMRF constitute the corticoreticular system. Considering that the reticulospinal tract is bilateral, influencing both contralateral and ipsilateral limb muscles, and that
26 reticulospinal cells receive direct cortical motor input, the reticulospinal tract is uniquely placed to make a strong contribution to motor recovery following cortical injury.
Many researchers are investigating the existence of alternate pathways through which the motor cortex can facilitate upper limb muscles (Bradnam and others, 2013;
Carmel and others, 2013; Herz and others, 2012; Liu and others, 2012; Rosenzweig and others, 2009; Rosenzweig and others, 2010; Tsai and others, 2011). However, this line of inquiry has been focused primarily on alternate pathways from the ipsilateral (uninjured) cortex (Rosenzweig and others, 2009; Rosenzweig and others, 2010) or interactions between the motor cortex and the red nucleus (Carmel and others, 2013; Herz and others,
2012; Liu and others, 2012; Tsai and others, 2011) or propriospinal system (Alstermark and others, 2007; Bradnam and others, 2013). The existence of a pathway between the motor cortex, the reticulospinal tract, and the motoneurons would allow for another alternate motor pathway by which the motor cortex could influence upper limb function.
Early tract tracing studies in the cat and monkey revealed that corticoreticular projections originate from all areas of the motor cortex, although as seen in figure 2.4 there is a tendency for the premotor areas such as SMA and PMd to send more projections than M1
(Keizer and Kuypers, 1984; Keizer and Kuypers, 1989) in the contralateral cortex.
Physiological studies have also shown the existence of a corticoreticular pathway from
M1, SMA, and PMA (Kably and Drew, 1998). Despite the evidence that a corticoreticular pathway does exist, our understanding of its role in motor control within the intact and injured nervous system is very limited. Though many researchers acknowledge that this system may play an important role in motor recovery following
27
Figure 2.4 Schematic diagram modified from Keizer and Kuypers (1989) showing the location of neurons in the contralateral macaque cortex that project to the spinal cord and/or brainstem. The authors injected diamindo yellow unilaterally in the spinal cord and fast blue in the reticular formation of macaques. They then studied the contralateral cortex to identify yellow neurons (corticospinal neurons), blue neurons (corticobulbar neurons – to reticular formation or facial nucleus), or green neurons (corticospinal neurons with collateral branches to the reticular formation). The authors identified that there were a greater number of projections from the premotor cortex (PMA and SMA) to the reticular formation than from M1 to the reticular formation.
cortical injury (Kaeser and others, 2010; Umeda and others, 2010), our limited understanding of the corticoreticular pathway prevents us from realizing the significance that changes within the corticoreticular system may have on motor recovery following neurological injury. To this end, the following experiments put forth in this dissertation are aimed to provide some clarity and detail about the organization of the corticoreticular pathway from SMA in the non-human primate.
Both anatomical and physiological studies have found evidence that the PMRF receives motor cortical information via two routes: (1) some primary corticospinal fibers 28 send collateral branches to the PMRF, and (2) corticoreticular neurons that arise from cortical motor areas and terminate in the PMRF but not in the spinal cord (Kably and
Drew, 1998; Keizer and Kuypers, 1984; Keizer and Kuypers, 1989; Matsuyama and
Drew, 1997; Newman and others, 1989; Rho and others, 1997). M1 sends more corticospinal collaterals to the PMRF compared to the number of corticoreticular neurons that travel from M1 and terminate in the PMRF (Kably and Drew, 1998). The opposite is true for SMA and PMA, where these areas send more corticoreticular neurons to the
PMRF compared to corticospinal collaterals (Kably and Drew, 1998). Anatomical evidence also suggests that this is the case with most corticospinal collaterals arising from M1 and corticoreticular neurons arising from the premotor areas (Keizer and
Kuypers, 1984; Keizer and Kuypers, 1989). This anatomical evidence however is incomplete as it is only described for projections from the contralateral cortex not the ipsilateral cortex. Given that SMA and PMA mainly function in movement preparation, this appears to suggest that M1 sends a copy of the movement as it is about to happen (as the collaterals are carrying the same message to the muscles and the PMRF at the same time), whereas SMA and PMA send information regarding upcoming movement prior to the movement occurring (as corticoreticular neurons send information solely to the
PMRF and not to the muscles). Thus, projections from SMA and PMA appear to be more strongly related to APAs whereas projections from M1 appear to be more related to dynamic postural adjustments.
29
Conclusion
The presence of a pathway from the motor cortex to spinal alpha motoneurons via the PMRF and reticulospinal tract (corticoreticulospinal pathway) is an understudied motor pathway through which both upper limbs can be activated due to the bilateral nature of reticulospinal projections. This pathway appears to have a significant role in postural control during movements, both in preparation for movement and whilst the movement is underway. Of most interest is the role that this pathway may play in the postural control of the upper limb while reaching movements and distal hand movements are being performed. Both the reticulospinal tract and SMA have been shown to be important in postural control, however these two structures have not yet been studied at the same time in order to ascertain if interactions between these two structures contributes to motor control of the upper limb.
The significance of setting and controlling postural movements during upper limb tasks is often overlooked because of the subtle nature of this task, however following neurological injury, it is this lack of proximal stability that contributes to continued disability and decreased quality of life. Of all of the three cortical motor areas discussed, the motor responses associated with SMA neuron activity share some similarities with the motor responses observed with reticulospinal activity. Firstly, SMA and reticulospinal neurons both elicit responses mostly in muscles of the proximal upper limb around the shoulder and trunk. Secondly, a number of SMA and reticulospinal neurons show increased activity during periods of movement preparation when APAs are initiated.
30
Thirdly, of all the cortical motor areas, SMA neurons are most likely to produce bilateral and ipsilateral muscle responses, which is also a characteristic of the reticulospinal tract.
These similarities in upper limb responses associated with SMA and reticulospinal activity indicate that a corticoreticulospinal pathway between SMA, reticulospinal neurons, and the spinal cord may exist and that it plays a role in the postural control of the upper limb during limb movements. This pathway provides an avenue by which recovery of upper limb function could be achieved following neurological injury.
However a greater understanding of this corticoreticulospinal pathway is first needed so that rehabilitation strategies that will recruit this pathway can be designed.
In the next chapter, the organization of the reticulospinal tract in the non-human primate is detailed. Questions of the laterality of reticulospinal projections and the locations of reticulospinal cells within the PMRF are addressed. Additionally, this chapter also details reticulospinal cell size, and the role that the small and large reticulospinal cells may play in the motor control of the upper limb is discussed. In chapter four, the pattern of muscle recruitment in the upper limbs associated with activity in three cortical motor areas (M1, PMd, and SMA) is investigated. In particular, a comparison of each of the three areas within the same subject is made to address whether
SMA is more likely to demonstrate a bilateral or ipsilateral pattern of muscle recruitment in the upper limbs compared with other cortical areas. Finally in chapter five, the organization of corticoreticular projections from SMA to the PMRF is described. In this chapter, questions regarding the laterality and distribution of corticoreticular projections are addressed and evidence of possible connections between SMA and reticulospinal
31 cells is presented. The studies detailed in the following chapters provide evidence that corticoreticular projections exist between SMA and the reticulospinal cells in the non- human primate and that SMA elicits muscle responses in proximal muscles in the ipsilateral and contralateral upper limbs, muscles that have previously been shown to be recruited following reticulospinal activity. Demonstrating the existence of a corticoreticulospinal pathway in a primate model is an important step to understanding the organization of this pathway within humans. The evidence presented in this dissertation can guide further research within primates and humans to understand how this pathway may control upper limb function in the intact nervous system, and ultimately within the injured nervous system.
32
Chapter 3 Anatomical organization of cells of the reticulospinal tract in the
pontomedullary reticular formation of the non-human primate
The pontomedullary reticular formation (PMRF) is composed of a number of loosely organized nuclei located medially in the pons and medulla (Brodal, 1956;
Newman, 1985a; Newman, 1985b; Valverde, 1961). Many cells within the PMRF have fibers that project to the spinal cord as the reticulospinal tract (Matsuyama and others,
1993; Matsuyama and others, 1999; Matsuyama and others, 1997; Peterson and others,
1975; Peterson and others, 1979; Torvik and Brodal, 1957). These reticulospinal cells send projections to various levels of the spinal cord where they influence motor outputs that are concerned with the control of posture and locomotion (Mori and others, 1998;
Prentice and Drew, 2001; Schepens and Drew, 2004; Stapley and Drew, 2009; Wannier and others, 1998). Research has also shown that the reticulospinal tract is involved with upper limb (UL) function (Buford and Davidson, 2004; Davidson and Buford, 2006;
Schepens and Drew, 2004; Schepens and Drew, 2006). However, the details concerning how the reticulospinal system contributes to UL stability and movement are still not clearly understood.
Physiological studies in cats and non-human primates (NHP) have demonstrated that stimulation of the reticulospinal tract leads to activation of a number of arm and
33 trunk muscles (Buford and Davidson, 2004; Davidson and Buford, 2006; Herbert and others, 2010; Peterson and others, 1979; Peterson and others, 1978). There is typically a bilateral pattern observed following stimulation, with arm flexors facilitated on the side of the stimulation and arm extensors facilitated on the side opposite the stimulation
(Buford and Davidson, 2004; Davidson and Buford, 2006; Herbert and others, 2010).
Although some recent work has uncovered evidence that the reticulospinal tract may contribute to hand function in the NHP (Baker, 2011; Riddle and others, 2009;
Soteropoulos and others, 2012), most studies have demonstrated that the muscles most likely to be activated following reticulospinal activity are those in the proximal arm involved with moving the entire upper limb or stabilizing the arm during hand movements (Davidson and Buford, 2006; Herbert and others, 2010; Schepens and Drew,
2003). The mechanism by which these responses are elicited however, are still not clear.
Stimulation studies indicate that the projections from the reticulospinal tract to the spinal cord are strongly ipsilateral (Peterson and others, 1979; Peterson and others, 1978).
Anatomical studies describing the structure and organization of the reticulospinal tract in different animal models – rodents (Newman, 1985a; Newman, 1985b; Zemlan and Pfaff,
1979), opossums (Beran and Martin, 1971; Martin and Dom, 1971; Martin and others,
1979), and cats (Matsuyama and others, 1993; Matsuyama and others, 1999; Matsuyama and others, 1997) – also indicate that the reticulospinal tract projects primarily ipsilaterally from the medial reticular formation in the brainstem to the ventral funiculus of the spinal cord. Currently it is thought that the bilateral motor outputs observed in the
UL’s following reticulospinal tract activity results chiefly through the influence of
34 reticulospinal activity on local spinal circuitry involving commissural interneurons
(Bannatyne and others, 2003; Jankowska and others, 2003) and not from the smaller number of contralaterally descending reticulospinal projections.
These studies, however, have only been conducted in lower order animal models and it is not clear if the reticulospinal tract in primates which influences UL motor function is organized in a similar way. Does the reticulospinal system in the NHP project bilaterally from the PMRF to the spinal cord, or is it predominantly an ipsilateral tract, as in lower order animals? Discovering the answer to this question will help us understand how this motor system may be recruited to assist with UL recovery following nervous system injuries such as stroke and spinal cord injury. At this time only two published studies have addressed the organization of the reticulospinal tract in NHP using tract tracers (Kneisley and others, 1978; Sakai and others, 2009). Unfortunately, these studies present conflicting evidence regarding the nature of the reticulospinal tract leaving us no closer to understanding details of its organization in the NHP. Kniesley and colleagues
(Kneisley and others, 1978) provided a qualitative description of the reticulospinal tract in the NHP, which described the reticulospinal tract as being predominantly ipsilateral, similar to the findings of the reticulospinal tract in other animal models. However, Sakai and colleagues (Sakai and others, 2009) provided quantitative data suggesting that the reticulospinal tract in the NHP is a bilateral tract with little difference between projections travelling ipsilaterally and contralaterally in the spinal cord, in contrast to reports of the reticulospinal tract in other animal models.
35
We conducted this study in order to describe the organization of the reticulospinal system that projects to the cervical cord in the NHP using well established quantification techniques (unbiased stereology). We used unilateral retrograde tract tracers injected into the cervical enlargement to label reticulospinal cells that projected to the cervical cord.
The purpose of this study was twofold: firstly, to detail if there was a difference in the distribution of reticulospinal cells throughout the PMRF both in terms of cell number and cell size; secondly, to determine if reticulospinal cells are located ipsilaterally and contralaterally with regards to the retrograde injection.
Methods
Three adult (3-5 years old) female subjects (M. fascicularis) were used for this study, which was conducted in compliance with protocols approved by The Ohio State
University Institutional Animal Care and Use Committee. Prior to the surgery for the retrograde injections, subjects received anterograde tracer injections in bilateral supplementary motor areas as part of a different study and those methods and results are not described here.
Surgical Procedures
All surgical procedures were carried out using sterile techniques. Subjects were anesthetized with ketamine hydrochloride (5 mg/kg, IM). A preoperative dose of antibiotics (florfenicol; Nuflors - 0.45 ml SQ) and dextramethasone (Solu-Delta Cortef –
36
0.7 ml IM) were administered. Subjects were intubated and kept on isoflurane anesthesia
(2L/min) throughout the surgery. Heart rate, respirations, oxygen saturation, and body temperature were monitored throughout the surgery to ensure subjects remained hemodynamically stable. Subjects were placed in prone in a stereotaxic frame (Kopf instruments, Tujunga, CA) and the cervical vertebrae exposed. A laminectomy was performed at the C5-C6 level and the dura was removed, exposing the dorsal cord surface.
A 10 μl Hamilton microsyringe with a 28 gauge needle (Hamilton, Reno, NV) was loaded with 1% cholera toxin subunit B (CTb) (Sigma-Aldrich, St Louis, MO). The syringe was placed in a stereotaxic carrier and positioned on a custom made dorsal attachment allowing the injections to reach across the entire medio-lateral extent of the spinal cord. To avoid tracer spread to the other side of the cord, the most medial injections were made 0.75 mm lateral to midline. Unilateral pressure injections of CTb were made at the C5-6 spinal levels in rows from rostral to caudal, with each row separated from the next by 1 mm. In each row, injections were made from medial to lateral at 1mm intervals. At most sites, 0.2 uL was injected 3.5 mm from the cord surface, followed by another 0.2 μL one minute later. After two minutes the needle was raised one mm and another 0.3 uL was injected for a total of 0.7 uL. To avoid injecting white matter regions, at the most lateral sites, only one injection of 0.3 uL was made 2.5 mm from the surface. Once all CTb injections were made, unilateral pressure injections of 3% fluorogold (FG) (Fluorochrome, Denver, CO) in triton (Tx-100) (Jefferson and others, 2011) were made on the opposite side of the cord using then same protocol outlined above. Gel film (Pharmacia and Upjohn Co, Kalamazoo, MI) and gel foam
37
(Surgifoam; Ethicon inc., Somerville, NJ) were then placed over the cord and into the vertebral deficit and the incision was closed. Subjects recovered and received postoperative analgesia with bupenorphine (0.15 ml IM) and metacam (0.2 ml PO) and a prophylactic course of antibiotics (florfenicol – 0.45 ml SQ) to manage post-operative pain and limit the risk of post-operative infection.
Tissue Processing
Two weeks following surgery, subjects were anesthetized with sodium pentobarbital (Nembutal - 3 ml IV) and transcardially perfused with 0.1 M phosphate buffered saline (PBS) followed by 4% paraformaldyhyde with periodate-lysine (PLP) in phosphate buffer (PB). Tissues were then perfused with solutions of 10% sucrose in PB, and finally 30% sucrose in PB. The brain and cervical spinal cord were then removed and post fixed overnight in 4% PLP. The tissue was then placed in a 30% sucrose solution until it was frozen using isopentane and dry ice (Rosene and others, 1986) and stored at -70° C. The segment of the spinal cord where tracer injections were made was cut longitudinally at 50 uM using a Leica sm2000R sliding microtome (Leica
Microsystems, Buffalo Grove, IL). Brainstem tissue was cut in 50 uM coronal sections and collected in 10 series for further staining. One series was immediately mounted and stained for cresyl violet for cytoarchitectural comparison.
38
Immunohistochemistry
Spinal Cord
Several spinal cord sections for each subject were mounted with immumount
(Thermo scientific, Waltham, MA) to visualize the FG injection sites. Another set of cord sections was then stained for using antibodies raised against CTb and/or FG for light microscopy viewing of the tracer injection sites. Sections were rinsed in PBS and then incubated in 30% H2O2 in methanol for 15 minutes to eliminate endogenous peroxidase.
Following further PBS rinses, sections were incubated in a blocking solution for one hour
- 5% normal horse serum (NHS)/1% bovine serum albumin (BSA)/1% Tx-100 in PBS.
To identify FG, sections were incubated in rabbit anti-FG (1:10,000; Fluorochrome,
Denver, CO) overnight at room temperature (RT), rinsed in PBS and incubated for two hours in biotinylated horse anti-rabbit secondary antibody (1:600; Vector Labs,
Burlingame, CA) followed by avidin-biotin complex (ABC Elite) (Vector Labs,
Burlingame, CA) for one hour. The complex was then visualized with H2O2 and 3,3’- diaminobenzidine (DAB; Vector Labs, Burlingame, CA) for five minutes. Following several PBS rinses, sections were then incubated in goat anti-CTb (1:40,000; List
Biological Labs, Campbell, CA) for one hour at RT and then at 4°C overnight. Sections were then rinsed in PBS and incubated in a biotinylated horse anti-goat secondary antibody (1:600; Vector Labs, Burlingame, CA) for two hours followed by ABC complex for one hour and PBS rinse. The CTb complex was visualized by incubating sections in
H2O2 with Vector Very Intense Purple (V-VIP; Vector Labs, Burlingame, CA) as the chromagen, for two and a half minutes. Sections were mounted on subbed slides and
39
Figure 3.1 Photomicrographs of the retrograde tracer injections sites for each subject. (A) shows the FG injection sites for subject A, (B) shows the CTb injection sites for subject G, and (C) shows the CTb injection sites for subject R. The dashed line on each photomicrograph represents the section midline. Analysis of sections spanning the entire extent of the cervical spinal cord, confirmed all injections were unilateral with no spread to the opposite side of the cord. Scale bar is 2 mm.
40 coverslipped for light microscopy. Figure 3.1 shows the spinal cord injection sites for each subject. None of the injections had evidence of spread to the contralateral side of the spinal cord. Antibody specificity controls were done by following the above protocols but eliminating the primary antibody step. This resulted in no cell labeling with
CTb (when the goat anti-CTb step was eliminated) or FG (when the rabbit anti-FG step was eliminated).
Brainstem sections
Series of equally spaced 50 um thick sections spanning the full rostrocaudal extent of the brainsem were stained for CTb and/or FG to identify labeled reticulospinal cells. The sections were stained according to the protocol outlined above, except that for some sets the CTb or FG was visualized using Vector blue as the chromogen (Vector
Labs, Burlingame, CA). In place of peroxidase ABC, these sections were incubated in an
ABC alkaline phosphatase (ABC-AP; Vector Labs, Burlingame, CA) for an hour and then in vector blue solution (Vector Labs) for 20 minutes. Levamisole was added to the vector blue to eliminate endogenous alkaline phosphatase and decrease background staining.
To confirm the cytoarchitectural boundaries of the PMRF, one series of brainstem sections was stained with antibodies raised against the neurotransmitters, substance P and serotonin (5-HT). These sections were incubated for 48 hours at 4°C in goat anti 5-HT
(1:20,000; Immunostar, Hudson, WI) and rabbit anti-Substance P (1:10,000; EMD
Millipore, Waltham, MA). They were then rinsed and incubated first in a biotinylated
41 horse anti-goat secondary antibody (1:600; Vector Labs, Burlingame, CA), followed by
ABC, and H2O2 with V-VIP to label 5-HT. Sections were then incubated in biotinylated horse anti-rabbit secondary antibody (1:600; Vector Labs, Burlingame, CA) for two hours, followed again by ABC, and DAB used to visualize the substance P.
Data Analysis
Boundaries for the PMRF nuclei
Using CV, Substance P and 5-HT stained brainstem sections in combination with a stereotaxic atlas (Paxinos and others, 2000), we identified the boundaries of the PMRF.
These boundaries allowed us to define the PMRF region of interest (ROI) in each of the
CTb and FG labeled sections that were used for cell counting. Cell laterality (numbers of cells were located in the ipsilateral or contralateral PMRF) is always reported with respect to the side of the tracer injection in the cord.
The PMRF consists of three major nuclei – nucleus reticularis pontis oralis (PnO), nucleus reticularis pontis caudalis (PnC) and nucleus reticularis gigantocellularis (Gi). In this study we focused on the regions of the PMRF (PnC and Gi) where we have seen pronounced electromyographic (EMG) activity in muscles of the UL following electrical stimulation (Davidson and Buford, 2006). We have also included the caudal part of PnO
(cPnO) as we do elicit some UL muscle activity when we stimulate the caudal part of this nucleus. We separated Gi into a rostral and caudal part, as studies from our lab have shown that UL effects originate mostly from the rostral parts of Gi (Buford and
Davidson, 2004; Davidson and Buford, 2006).
42
The rostral boundary for cPnO was marked by the decussation of the trochlear nerve and the caudal boundary was delineated by the caudal portion of the parabrachial nucleus. PnC begins rostrally with the caudal end of the parabrachial nucleus and ends caudally at the abducens nucleus where the genu of cranial nerve VII appears. The rGi begins rostrally at the caudal end of the abducens nucleus when the solitary nucleus appears. The appearance of the dorsal cochlear nucleus marked the end of rGi and the beginning of cGi. The medial border for all PMRF nuclei was the raphe nuclei. The lateral boundary of cPnO was the central tegmental tract and the motor nucleus of V. In
PnC, the lateral edge was bound by the motor nucleus of V rostrally, and the facial nucleus and nerve caudally. The facial and inferior olivary nuclei formed the lateral boundary of rGi. The inferior olive and central tegmental tract formed the lateral border of cGi.
The pyramids and medial lemniscus formed the ventral border for cPnO, PnC, and rGi. In cGi, the inferior olive alone formed the ventral border. The dorsal border in cPnO was formed by the superior cerebellar peduncle and parabrachial nucleus, and the dorsal boundary in PnC and rGi was formed by the abducens nucleus. In cGi the dorsal boundary was the dorsal paragigantocellularis.
Stereological Approach
Cells were counted in CTb and FG stained brainstem sections using brightfield microscopy with the optical fractionator method and systematic random sampling (Stereo
Investigator version 10; MBF Bioscience, Williston, VT). Each section examined was
43 separated from the next by 0.5 mm. Using a Nikon eclipse E800 microscope and a SPOT
RTse CCD camera (Diagnostic Instruments Inc., Sterling Heights, MI), the ROI was outlined for each section. The Stereo Investigator program placed unbiased virtual counting frames throughout the ROI. Each counting frame was 250 μM2, and each frame sat within a 750 μM2 sampling grid. Each counting frame was examined through the z axis using a 60x objective, and a marker was placed on the tops of nuclei of cells containing CTb or FG as they appeared within each frame. Cell size was determined at the same magnification by measuring the diameter of each labeled cell at the widest part of the soma. We calculated the coefficient of error (CE) for each subject to determine the precision of our estimates according to the formula outlined by West (West and others,
1991). For subject G, the CE was 0.058 using 13 sections for analysis, for subject R the
CE was 0.097 using 10 sections for analysis, and for subject A the CE was 0.043 using 17 sections for analysis. The mean section thickness was 30 μM.
Post microscopy analysis
For each subject, the ROI and markers for each section were saved and imported into Corel Draw X5 (Corel Corporation, Ottawa, ON, Canada). The ROI and markers for all sections within each individual PMRF nucleus were overlaid onto each other and a single nuclear ROI was created by tracing around all sections to form one contour for the ipsilateral and contralateral regions for each nucleus. A goal of this study was to detail the distribution of the cells within each rostrocaudal level of the PMRF nucleus. To do this, the ipsilateral and contralateral regions of each nucleus were divided into six parts.
44
The coronal plane was divided into a dorsal and ventral zone, and the sagittal plane into a lateral, intermediate and medial zone, giving six zones – dorsolateral (DL), dorsointermediate (DI), dorsomedial (DM), ventromedial (VM), ventrointermediate (VI), and ventrolateral (VL). The number of cells in each zone was counted and expressed as the proportion of total cells found in the ipsilateral or contralateral nucleus.
Figures with photomicrographs were constructed using Adobe Photoshop CS 5.1
(Adobe Systems Incorporated, San Jose, CA). If brightness and contrast was adjusted to make details of the photomicrographs clearer, the adjustments were applied equally to the entire image.
Statistical Analysis
All data was entered into Microsoft Excel 2007. Minitab version 16 (Minitab inc,
State College, PA) was used for all statistical calculations. Kruskal-Wallis analyses were used to identify differences in the distribution of cells based on laterality (in relation to the retrograde injection site) and the rostro-caudal extent of the PMRF. Mann Whitney tests were performed to compare cell distribution within and between individual PMRF regions. Chi square tests were run to analyze differences in the distribution of cells throughout each PMRF nucleus, and to analyze differences in the distribution of large and giant cells throughout each PMRF nucleus. A Tukey’s correction was made when required to account for multiple comparisons. Significance was set at p< 0.05.
45
Results
Distribution of reticulospinal cells in the PMRF
Subject PMRF nucleus Total cPnO PnC rGi cGi ipsi contra ipsi contra ipsi contra ipsi contra ipsi contra A 8 4 45 3 55 13 33 17 141 37 G 8 2 57 7 70 11 28 7 163 27 R 4 5 29 9 33 14 9 6 75 34 Total 20 11 131 19 158 38 70 30 Table 3.1 Total number of objects (top of labeled reticulospinal nuclei) counted in each PMRF nucleus for each subject. As shown in the table, there were more objects in the ipsilateral PMRF nuclei compared to the contralateral nuclei, especially in PnC and rGi.
Retrogradely labeled cells were seen throughout the rostro-caudal extent of the
PMRF (figure 2). Table 3.1 lists the total number of objects (cells) counted in each
PMRF region and table 3.2 lists the estimated number of cells in the delineated regions of
PMRF for each subject. On average there were an estimated 15,000 labeled cells throughout the PMRF per subject, which is about 1/5 of the number of cells estimated to form the corticospinal tract as report by He and colleagues (He and others, 1995). The median estimated number of cells for each section was 100 for subject A, 155 for subject
G, and 125 for subject R. Analysis showed that there was no statistical difference between subjects based on the estimated number of cells found per section (H = 2.00, p =
0.368). For each subject, cells were found throughout the PMRF from cPnO to cGi, although as shown in table 1 and table 2, for all subjects there were more cells found in
46
Subject PMRF nucleus Total cPnO PnC rGi cGi ipsi contra ipsi contra ipsi contra ipsi contra ipsi contra A 800 400 4500 300 5500 1300 3300 1700 14100 3700 G 800 200 5700 700 7000 1100 2800 700 16300 2700 R 400 500 2900 900 3410 1480 900 600 7610 3480 Total 2000 1100 13100 1900 15910 3880 7000 3000 Table 3.2 Estimated number of labeled reticulospinal cells found in each PMRF nucleus for each subject. There are more cells in the ipsilateral nuclei compared to the contralateral nuclei, especially in PnC and rGi.
PnC and rGi. Figure 3.3 shows that > 70% of all labeled cells were found in PnC and rGi
for all subjects. For each subject, cPnO had the fewest labeled cells and rGI had the
most. The difference in cell distribution throughout the PMRF nuclei was significant for
subject A (H = 10.84, p = 0.013) and subject G (H = 8.51, p = 0.037). Although in
subject R there were a greater estimated number of cells in PnC and rGi (median PnC
cells per section = 120, median rGi cells per section = 179) compared to cPnO (median =
45) and cGi (median = 75), this difference did not reach statistical significance. For
subject A and subject G, there were significantly more cells in PnC (subject A – median
= 100 cells per section; subject G – median = 155) and rGi (subject A – median = 130
cells per section; subject G – median = 195) compared to cPnO (subject A – median = 30,
W = 10.5, p = 0.03; subject G – median = 40, W = 6.0, p = 0.05).
47
Figure 3.2 Photomicrographs showing the distribution of labeled reticulospinal cells in each of the four PMRF nuclei. A – D show the ipsilateral and contralateral PMRF in each of the PMRF nuclei (A shows cPnO, B shows PnC, C shows rGi, and D shows cGi). The ipsilateral side is on the left and the contralateral side is on the right with the black dashed line showing midline. There are more cells in the ipsilateral PMRF compared to the contralateral in each of the nuclei. E – L show examples of vector blue AP-stained cells containing CTb from the ipsilateral and contralateral PMRF for each nucleus (E, F from cPnO; G, H from PnC; I, J from rGi; and K, L from cGi). Labeled cells with nuclei in the plane of the photomicrogaph are identified with arrowheads. Brown fiber staining is in axons that were labeled as part of a different experiment. A – D scale bar is 1000 μM. E – L scale bar is 100 μM.
48
Subject A Subject G
c PnO c PnO 6% 5% c Gi c Gi 23% 17%
PnC PnC 35% 32%
r Gi r Gi 39% 43%
Subject R
c PnO 5% c Gi 12%
PnC 38%
r Gi 45%
Figure 3.3 Chart illustrating the average proportion of reticulospinal cells in each nucleus for each subject. A majority (> 70 %) of the reticulospinal cells were located in PnC and rGi in each subject. In contrast, there were relatively few cells in cPnO (5-6 %).
As seen in figure 3.2, cells were located in the ipsilateral and contralateral PMRF.
Overall, there were more cells in the ipsilateral PMRF compared with the contralateral
PMRF in all three subjects (median number of ipsilateral cells per section - subject A =
80, subject G = 145, and subject R = 85; median number of contralateral cells per section
49
– subject A = 15, subject G = 20, and subject R = 35). This predominance for cells to be in the ipsilateral PMRF was significant for subject A (H = 14.83, p < 0.001) and subject
G (H = 14.35, p < 0.001). Although this predominance did not reach significance in subject R, there was a trend towards significance (H = 2.12, p = 0.145). This ipsilateral
Figure 3.4 Graph of the total estimated number of cells found ipsilaterally and contralaterally in each PMRF nucleus. The black square is for subject G, the gray diamond subject A, and the dark gray triangle subject R. In all subjects, more reticulospinal cells were found in the ipsilateral PMRF across all nuclei. In cPnO and cGi the distribution of cells was more bilateral than in PnC and rGi where there was a larger disparity between the number of cells in the ipsilateral and contralateral PMRF. For subject A and G the difference between the number of cells in the ipsilateral and contralateral PMRF was significant in PnC and rGi (p < 0.05 indicated by the red star).
50 predominance was most evident in PnC and rGi where for every subject > 70% of cells were located in the ipsilateral PMRF (figure 4). When comparisons were made within each nucleus, this difference in cell laterality was significant only in PnC (subject A - W
= 40.0, p = 0.012; subject G – W = 26.0, p = 0.03) and rGi (subject A - W = 40.0, p =
0.012; subject G – W = 26.0, p = 0.03). Although the difference was not statistically significant, figure 4 shows that for subject R the ipsilateral cell predominance was also larger in PnC and rGi compared to cPnO and cGi.
Within the PMRF nuclei the cells are often concentrated in different areas of the nucleus (figure 3.5-7). In light of this observation, the ipsilateral and contralateral regions of each PMRF nucleus were divided into six zones and the number of cells in each zone was counted. Figures 3.5-7 show the distribution of these cells in each nucleus of the ipsilateral and contralateral PMRF for each subject. The diagrams show how the cells are distributed throughout the six zones. When we analyzed the distribution of reticulospinal cells within the ipsilateral PMRF, there was a difference in the distribution of cells ventrally and dorsally for all subjects (subject A - χ2 = 1054.2, p < 0.001; subject
G - χ2 = 3045.2, p < 0.001; and subject R - χ2 = 167.1, p < 0.001). Analyzing each PMRF nucleus, the greatest difference in cell distribution was in PnC, where cells were located more ventrally, and in rGi, where cells were located more dorsally. In terms of medio- lateral distribution of cells in the ipsilateral PMRF, there were a greater number of cells in the medial and intermediate zones compared to the lateral zones for all subjects
(subject A - χ2 = 349.4, p < 0.001; subject G - χ2 = 1274.0, p < 0.001; and subject R - χ2 =
856.2, p < 0.001). Analyzing each nucleus, there is a rostro-caudal pattern in cell
51
52
Figure 3.5 Schematic diagrams showing the distribution of reticulospinal cells throughout each PMRF nucleus for subject A. A illustrates the grid used to divide the ipsilateral and contralateral region of each nucleus into six zones (as noted in the methods). In B, the schematic diagrams for each PMRF nucleus taken from the Szabo and Cowan atlas (Szabo and Cowan, 1984) are also shown to demonstrate where each nucleus sits within the brainstem. The landmarks used to determine the boundaries of each of the PMRF nuclei are described in detail in the methods. C shows an enlarged view of each nucleus showing the position of the labeled reticulospinal cells (outlined in orange) counted for subject A. The distribution of the cells throughout each nucleus is shown for both the ipsilateral and contralateral PMRF with the grid overlaid. As illustrated, there are significantly more cells in ipsilateral PnC and rGi than in the other nuclei in the ipsilateral and contralateral PMRF. More cells were located in the medial and intermediate zones of ipsilateral PnC and cGi than in the lateral zones. There was no such pattern in the contralateral PMRF or the ipsilateral cPnO and cGi. 52
53
Figure 3.6 Schematic diagrams showing the distribution of reticulospinal cells throughout each PMRF nucleus for subject G. As in figure 5, A illustrates the grid used to divide the ipsilateral and contralateral region of each nucleus into six zones. In B, the schematic diagrams for each PMRF nucleus taken from the Szabo and Cowan atlas (Szabo and Cowan, 1984) are also shown to demonstrate where each nucleus sits within the brainstem. C shows an enlarged view of each nucleus showing the position of the labeled reticulospinal cells (outlined in orange) counted for subject G. As with subject A, subject G also had more cells in ipsilateral PnC and rGi than in the other nuclei in the ipsilateral and contralateral PMRF. Significantly more cells were also located in the medial and intermediate zones of ipsilateral PnC and rGi than in the lateral zones. This was not the case for ipsilateral cPnO and cGi. There was no such pattern in the contralateral PMRF.
53
54
Figure 3.7 Schematic diagrams showing the distribution of reticulospinal cells throughout each PMRF nucleus for subject R. As in figure 5, A illustrates the grid used to divide the ipsilateral and contralateral region of each nucleus into six zones. In B, the schematic diagrams for each PMRF nucleus taken from the Szabo and Cowan atlas (Szabo and Cowan, 1984) are also shown to demonstrate where each nucleus sits within the brainstem. C shows an enlarged view of each nucleus showing the position of the labeled reticulospinal cells (outlined in orange) counted for subject R. As with subjects A and G, this subject also had more cells in ipsilateral PnC and rGi than in the other nuclei in the ipsilateral and contralateral PMRF. However this difference in cell laterality was not as significant as it was for the other two subjects. As with the other two subjects, significantly more cells were also located in the medial and intermediate zones of ipsilateral PnC and rGi than in the lateral zones. There was no such pattern in the contralateral PMRF or ipsilateral cPnO and cGi. 54 distribution with cells moving from a lateral to medial distribution as the location moves from rostral to caudal. For example, in cPnO a greater number of cells were found in the lateral zones, whereas in PnC a greater number of cells were found in the intermediate zones. In Gi a majority of the cells were found in the medial zones. In the contralateral
PMRF there was no clear pattern of cell distribution seen.
Soma size
Newman (Newman, 1985a; Newman, 1985b) provides the best characterization of reticular formation cells and categorized cells based on soma diameter. In this study, cells were categorized as either small (≤ 25 μM), medium (25.1 – 35 μM), large (35.1 –
54.9 μM), or giant (≥ 55 μM) according to criteria outlined by Newman (Newman,
1985a; Newman, 1985b). Figure 3.8 shows photomicrographs of cells from each of the
Figure 3.8 These photomicrographs show examples of cells from each of the four size groupings (small ≤ 25 μM, medium 25.1 – 35 μM, large 35.1 – 54.9 μM, and giant ≥ 55 μM). The arrows point to each cell in the photomicrographs. The line drawn on each cell represents the line from which the measurements of cell diameter were taken. As mentioned in methods, cell diameter was measured across the widest part of the soma.
55 four categories. Labeled reticulospinal cells ranged from 8.7 μM to 82.8 μM in soma diameter. The distribution of small, medium, large, and giant cells was not different between the ipsilateral and contralateral PMRF for all subjects, with around 50% of cells in the ipsilateral PMRF being large or giant and 50% of cells in the contralateral PMRF being large or giant. There was, however, a difference in the soma size throughout the rostro-caudal extent of the PMRF for all three subjects (subject A – χ2 = 209.6, p < 0.001; subject G – χ2 = 103.0, p < 0.001; subject R – χ2 = 231.7, p < 0.001). The median cell size in cPnO was 37.8 μM for subject A, 24.7 μM for subject G, and 12.4 μM for subject
R. In PnC the median cell size was 31.5 μM for subject A, 36.1 μM for subject G, and
24.7 μM for subject R. For all subjects the largest cells were found in rGi where the median cell size was 38.3 μM for subject A, 40.4 μM for subject G, and 30.6 μM for subject R. In cGi the median cell size was 29.8 μM for subject A, 31.4 μM for subject G, and 15.0 μM for subject R. There were more large cells in PnC and rGi compared to the cGi in all three subjects (subject A – 35% of cells in PnC and 37% of cells in rGi were large compared to 30% in cGi; subject G – 42% of cells in PnC and 38% of cells in rGi were large compared to 29%; and subject R – 19% of cells in PnC and 28% of cells in rGi were large compared to no cells in cGi). There were also more large cells in PnC and rGi compared to cPnO in subjects G and R (subject G – 30% of cells in cPnO were large; and subject R there were no large cells in cPnO). Giant cells were also preferentially located in PnC and rGi compared to cPnO and cGi where giant cells were very sparse.
Indeed, no giant cells were found in cPnO for any of the three subjects.
56
In the literature, there are variable reports of what constitutes a large or giant cell in the reticular formation (Lingenhohl and Friauf, 1992; Newman, 1985a; Newman,
1985b; Sakai and others, 2009). Cells that were ≥ 35 μM have been designated as giant neurons in some cases (Lingenhohl and Friauf, 1992), thus we combined large and giant cells into one group to make further comparisons of cell size in each PMRF region. Of the cells counted, most were either small or medium in size for subject A and R, while
780 (41.5%) for subject A and 210 (19.8%) for subject R were either large or giant. For subject G there were slightly more large and giant cells, 980 (51.6%), than small and medium cells. There was a significant difference in the proportion of large and giant cells in the different nuclei of the PMRF for each subject (subject A - χ2 = 130.1, p <
0.001; subject G - χ2 = 87.6, p < 0.001; and subject R - χ2 = 260.4, p < 0.001). Of the large and giant cells in all subjects, rGi had the highest number of cells (subject A - 50%; subject G – 47%; and subject R – 67%). This was then followed by PnC which had the second highest proportion of large and giant cells for all subjects (subject A - 23%; subject G – 37%; and subject R – 33%). Notably, as seen in figures 9-10, fewer large and giant cells were located in cPnO (subject A – 9%; subject G – 4%) and cGi (subject A –
18%; subject G – 13%). No large or giant cells were found in cPnO or cGi in subject R
(figure 3.11).
As seen in figures 3.9-11, a majority of large and giant cells were located in the ipsilateral PMRF. For all subjects, a majority of large and giant cells were in the ipsilateral PMRF - 62/79 (78%) in subject A; 83/98 (85%) in subject G; and 13/21 (62%)
57
58
Figure 3.9 Schematic diagrams showing the distribution of the large and giant reticulospinal cells throughout each PMRF nucleus for subject A. A illustrates the grid used to divide the ipsilateral and contralateral region of each nucleus into six zones (as noted in the methods). In B, the schematic diagrams for each nuclei taken from the Szabo and Cowan atlas (Szabo and Cowan, 1984) are also shown to demonstrate where each nucleus sits within the brainstem. The distribution of the large and giant cells throughout each nucleus is shown for both the ipsilateral and contralateral PMRF with the grid overlaid in C. As illustrated, there are significantly more large and giant cells in ipsilateral PnC and rGi than in the other nuclei in the ipsilateral and contralateral PMRF. Large and giant cells in ipsilateral PnC and rGi are located in the medial and intermediate zones.
58
59
Figure 3.10 Schematic diagrams showing the distribution of large and giant reticulospinal cells throughout each PMRF nucleus for subject G. As with figure 9, the grid used to divide each nucleus into six zones is shown in A. In B, the schematic diagrams for each nuclei taken from the Szabo and Cowan atlas (Szabo and Cowan, 1984) are also shown to demonstrate where each nucleus sits within the brainstem. In C, the distribution of the large and giant cells throughout each nucleus is shown for both the ipsilateral and contralateral PMRF with the grid overlaid. As illustrated, there are significantly more large and giant cells in ipsilateral PnC and rGi than in the other nuclei in the ipsilateral and contralateral PMRF. Large and giant cells in ipsilateral PnC and rGi are located in the medial and intermediate zones. In ipsilateral PnC, the large and giant cells were also located more ventrally, however the distribution in ipsilateral rGi was more even between the dorsal and ventral portions of the nucleus. 59
60
Figure 3.11 Schematic diagrams showing the distribution of the large and giant reticulospinal cells throughout each PMRF nucleus for subject R. Unlike subject A and G, subject R did not have any large or giant cells in cPnO and cGi therefore those nuclei are not presented in this figure. A shows the grid once again that was used to divide each nucleus into six zones. In B, the schematic diagrams for PnC and rGi taken from the Szabo and Cowan atlas (Szabo and Cowan, 1984) are shown to demonstrate where each nucleus sits within the brainstem. In C, the distribution of the large and giant cells throughout PnC and rGi is shown for both the ipsilateral and contralateral PMRF with the grid overlaid. As illustrated, large and giant cells in ipsilateral PnC and rGi are located in the medial and intermediate zones.
60 in subject R. This was seen in all nuclei for all subjects. However, it was particularly pronounced in PnC where 94% of large and giant cells in subject A, 89% in subject G, and 71% in subject R were in the ipsilateral PMRF. In rGi this difference between the proportion of large and giant cells in the ipsilateral and contralateral PMRF was also marked with 77% of large and giant cells in subject A, 80% in subject G, and 57% in subject R were located in the ipsilateral PMRF. The laterality of giant cells alone was even more striking, with 100% of giant cells in PnC and cGi in all subjects being located in the ipsilateral PMRF. In rGi, 80% of giant cells were in the ipsilateral PMRF for subject G and 67% of giant cells were in the ipsilateral PMRF for subject A.
Most of the large and giant cells were located within close proximity of each other in PnC and rGi, whereas they are more sparsely dispersed in cPnO and cGi. In the ipsilateral PMRF, large and giant reticulospinal cells were located in the intermediate and medial zones of PnC and rGi compared to the lateral zones for all three subjects (subject
A - χ2 = 297.6, p < 0.001; subject G - χ2 = 1282.0, p < 0.001; and subject R - χ2 = 72.6, p
< 0.001). The distribution of cells dorsally and ventrally was somewhat mixed with cells located ventrally in PnC for subjects A and G (subject A - χ2 = 820.2, p < 0.001 and subject G - χ2 = 1764.1, p < 0.001). However, there was no distinct distribution of cells dorso-ventrally in the other nuclei. In the contralateral PMRF, large and giant reticulospinal cells were located ventrally in PnC and rGi in subject A and G (subject A -
χ2 = 53.8, p < 0.001 and subject G - χ2 = 730.8, p < 0.001). In terms of medio-lateral distribution, as in the ipsilateral PMRF large and giant cells in PnC and rGi were located predominantly in the medial and intermediate zones. However, in cGi large and giant
61 cells were found most often in the lateral portions of the PMRF for both subject A (χ2 =
1181.2, p < 0.001) and subject G (χ2 = 1726.2, p < 0.001).
Discussion
This study is the first to document the full distribution pattern of reticulospinal cells from all regions of the PMRF that project to the gray matter of the NHP cervical spinal cord. Our results confirm and extend the findings of Kneisley and colleagues
(Kneisley and others, 1978) that although the reticulospinal tract extends both ipsilateral and contralateral projections to the cervical spinal cord, there is a strong preponderance for projections that extend from the PMRF to the ipsilateral spinal cord. Despite clear profound functional differences in forelimb function between species, these results are consistent with data from rat, opossum and cat (Beran and Martin, 1971; Martin and
Dom, 1971; Martin and others, 1979; Matsuyama and others, 1993; Matsuyama and others, 1999; Matsuyama and others, 1997; Newman, 1985a; Newman, 1985b; Zemlan and Pfaff, 1979) suggesting that the organization of the reticulospinal tract is phylogenetically conserved. We have also provided the first quantitative survey to confirm that in the NHP the majority of reticulospinal cells are located in PnC and rostral
Gi, and that spinal projections arise mostly from the medial and intermediate zones of these nuclei. In addition, we have demonstrated that the majority of large and giant reticulospinal cells are found in PnC and rostral Gi compared to the other PMRF nuclei, while small and medium sized neurons are distributed throughout both the ipsilateral and
62 contralateral reticular formation. Together these findings demonstrate that larger cells from the rostral and medial regions of the ipsilateral PMRF constitute the majority of direct reticulospinal projections, while a smaller, but important proportion of reticulospinal projections also exist, and these arise largely from contralateral and ventromedial regions and are more prevalent in the caudal PMRF.
PnC and rGi are the main source of reticulospinal projections
Although the PMRF consists of a number of nuclei, activity in the two nuclei that lie in the caudal pons and rostral medulla (PnC and rGi) have been shown to relate most strongly with muscle activity in the UL (Buford and Davidson, 2004; Davidson and
Buford, 2006; Schepens and Drew, 2003; Schepens and Drew, 2004). Anatomically, it has also been shown that a majority of the reticulospinal cells originate from these nuclei as well (Beran and Martin, 1971; Newman, 1985a; Newman, 1985b). In this study we have confirmed that this level of structural organization also exists in the NHP. We have also revealed that of these cells in PnC and rGi a large proportion of them are large or giant in size (≥ 35 µM), this is also in line with earlier work in other animal models
(Lingenhohl and Friauf, 1992; Newman, 1985a; Newman, 1985b). These nuclei have also been shown to be important in eliciting postural responses, with the giant cells of
PnC being an integral part of the circuitry of the acoustic startle reflex (Lingenhohl and
Friauf, 1992; Yeomans and Frankland, 1995). The presence of a large number of large and giant reticulospinal cells in PnC and rGi has functional implications which, as
63 discussed below, suggest that the reticulospinal tract is important in the control of trunk and UL posture during reaching.
It is also important to note that the presence of cells in the intermediate and medial zones of PnC and rGi is notable mostly for the fact that previous studies have shown that many of the cerebellar and cortical motor inputs to the reticular formation are located in the medial regions of these nuclei (Homma and others, 1995; Matsuyama and
Drew, 1997; Newman and others, 1989). Thus, it appears that the large and giant reticulospinal cells are located in areas of the PMRF that receive significant input from other motor systems involved in movement and postural control. This may provide evidence as to the functional importance of these cells.
The reticulospinal tract projects predominantly ipsilaterally
Our finding that reticulospinal cells project mainly ipsilaterally to the cervical spinal cord is in contrast to our own earlier published work (Sakai and others, 2009). We suspect that this is due to changes that we made to the methodological design of this study. First, on re-examination of the retrograde injection sites from specimens used for the previous paper we found that although we attempted to make the tracer injections unilaterally, there was occasional spread of the tracer injections to the other side of the cord. In the subjects where the tracer was not always confined to one side of the spinal cord it is impossible to accurately ascertain which cells sent projections ipsilaterally and which cells sent projections contralaterally. To prevent this from occurring in this study, our most medial injection site was 0.75 mm lateral to midline in contrast to the 0.5 mm
64 used in the earlier paper. Second, although our earlier paper was the first to quantitatively describe the reticulospinal system in the NHP, modern stereological sampling methods were not used. In the present study we used unbiased sampling methods and counted nuclei within labeled cells rather than counting labeled soma profiles. This approach provided an accurate estimate of cell numbers throughout the entire PMRF instead of unintentional selection of regions of the PMRF where there was a high concentration of cells or preponderance of larger cells which are more likely to be overcounted. Thirdly, the previous study (Sakai and others, 2009) used parasagittal sections to allow visualization of the full rostrocaudal distribution of PMRF nuclei. This has two disadvantages for the type of comparison done here. Specifically, this orientation makes clear identification of the relationship of labeled cells to the brainstem midline difficult, and it also precludes ensuring that the ipsilateral and contralateral sections are stained identically, as they would be on separate tissue sections stained in different staining wells. Thus, we are confident that the current results provide a more definitive map of the distribution of reticulospinal tract cells in the NHP.
Although our findings regarding the laterality of reticulospinal cells was not in agreement with earlier work published by Sakai and colleagues, the description of reticulospinal cells in terms of cell size and the distribution throughout the PMRF nuclei was similar. In agreement with Sakai, our analysis showed that cells varied in size from small to giant, and that most of the giant cells were located in PnC and Gi. In addition we were also able to confirm Sakai’s finding that most of the reticulospinal cells are located in Gi and PnC compared to cPnO where cells were sparsely located.
65
The predominantly ipsilateral organization of the reticulospinal tract in the NHP suggests that the bilateral effects found following reticulospinal tract stimulation in the
NHP may result predominantly from influences of the reticulospinal tract on local spinal circuitry as demonstrated in the cat by Jankowska’s group (Bannatyne and others, 2003;
Jankowska and others, 2003). In cats, the reticulospinal tract influences bilateral limb muscle activity via synapses that the fibers make on commissural interneurons in the spinal cord (Bannatyne and others, 2003; Jankowska and others, 2003). Our finding that the reticulospinal tract is also predominantly ipsilateral in NHP, (Matsuyama and others,
1993; Matsuyama and others, 1999; Matsuyama and others, 1997), suggests that the bilateral motor effects observed following PMRF activity (Buford and Davidson, 2004;
Davidson and Buford, 2006; Herbert and others, 2010) may be explained by similar circuitry. It has been hypothesized that the bilateral motor outputs observed may be mediated by reticulospinal cells that project bilaterally (Peterson and others, 1975). Our approach precluded direct documentation of bilateral projections, but future studies using dual retrograde tracers with matching transport characteristics (Conte and others, 2009a;
Conte and others, 2009b) are planned as a followup to this work. In addition, the use of transsynaptic tracers to identify the entire reticulospinal circuit from the PMRF to muscles in the UL, which has been done in the NHP to trace corticospinal circuits (Dum and Strick, 2013), could be used to ascertain whether bilateral motor outputs are elicited through spinal interneurons in the NHP.
Limitations
66
One feature of the data presented here is the low number of cells observed in one subject (subject R). Although intersubject variability in tract tracing studies is common due to subject differences in the uptake and transport of tracers, in the case of subject R this variability was compounded by the fact that tissue fixation was not optimal. As a result, subject R’s tissue was very fragile and subsequent staining was limited due to this.
The low number of cells observed in this subject also made it difficult to reach statistical significance during data analysis. Thus, although the same trends of cell distribution that were observed in subject A and G were found in subject R (more cells were located in the ipsilateral PMRF, most cells were found in PnC and rGi, and large and giant cells were mostly found in PnC and rGi), they never reached statistical significance in subject R due to the low cell counts.
Functional implications
Henneman and colleagues (Henneman and others, 1965) concluded that there is a correlation between the size of a cell body and the size of its axon, with large cell bodies having large axons and thus faster conduction velocities. Evarts (Evarts, 1965; Fromm and Evarts, 1981) demonstrated that, in the corticospinal tract, large axons exhibit maximal discharge activity during large fast movements whereas small axons show maximal activity during slow, small movements. Similar attributes have been found in alpha motoneurons (Henneman and others, 1965). In terms of the reticulospinal system, work by Peterson has demonstrated that some reticulospinal cells can have conduction velocities of up to 101 m/sec (Peterson and others, 1975), similar to conduction velocities
67 from larger corticospinal axons found by Evarts (Evarts, 1965). Based on this work on the relationship between cell size, axon diameter, and conduction velocity, we can conclude that the large and giant reticulospinal cells have larger axons and thus, faster conduction velocities than smaller reticulospinal cells.
Given these tenets, Evarts work would lead us to conclude that the largest of the reticulospinal cells are most likely the cells which are involved in producing rapid movement of the largest of the UL and trunk muscles. Given that the UL muscles that are most associated with reticulospinal activity are postural and stability muscles around the trunk and shoulder girdle (Davidson and Buford, 2006; Herbert and others, 2010) these large and giant reticulospinal cells are primed to produce fast postural adjustments associated with perturbations that accompany reaching, especially toward distant objects
(Castiello and others, 1992; Gentilucci and others, 1992; Huang and Brown, 2013). The smaller reticulospinal cells, due to their slower conduction velocities are more likely to be involved with postural adjustments in preparation for movement. Such preparatory functions often involve small movements and slow, prolonged patterns of muscle activity and, as Evarts work has shown, cells with smaller axons are the most likely cells to be involved with eliciting such movements (Evarts, 1968; Evarts, 1969).
Additional studies indicate that reticulospinal cells receive not only multiple sensory inputs but also cerebellar inputs related to posture and balance (Eccles and others, 1975). This work suggests that these large reticulospinal cells may be involved in postural adjustments to ensure trunk stability in anticipation of movement and during reaching movements. For example, Eccles and colleagues found that the fastigial nucleus
68 of the cerebellum (which is involved in postural control) sends inputs preferentially to reticulospinal cells with large axons (Eccles and others, 1975). Thus the presence of large and giant cells in the ipsilateral PnC and rGi indicate that these nuclei are positioned as the primary reticular nuclei involved in eliciting the fast and sometimes large postural adjustments that maintain stability for the performance of more distal UL movements.
Conclusion
This study has demonstrated that the reticulospinal tract in the NHP is predominantly an ipsilateral tract that contains cells of varying sizes from small to giant.
The majority of these cells are located in the medial and intermediate portions of the
PMRF, areas where there are numerous of cortical motor and cerebellar inputs to the
PMRF. The anatomical organization of the reticulospinal tract is very similar to that of the cat, as are the previously published physiological findings related to reticulospinal activity in the NHP. The organizational and functional similarities in the reticulospinal system in the cat and NHP lead us to conclude that in the NHP, as in the cat, the bilateral motor responses elicited by reticulospinal activity are most likely due to local spinal circuitry. If this is the case, then following neurological injury changes in the reticulospinal tract that could facilitate functional recovery most likely occur at the level of the spinal cord. Here, reticulospinal inputs to spinal interneurons could be altered
(through either axonal sprouting or pruning) that would then lead to a difference in the
69 muscle recruitment pattern and may play a role in the recovery of upper limb function following injury.
70
Chapter 4 Recruitment of ipsilateral and contralateral upper limb muscles following
stimulation of the cortical motor areas in the monkey.
In primates, projections from the three largest motor areas in the frontal cortex, primary motor cortex (M1), supplementary motor area (SMA), and dorsal premotor area
(PMd), make up a significant proportion (~70%) of corticospinal tract fibers (Brinkman and Kuypers, 1973; He and others, 1995; Kuypers, 1960; Lawrence and Kuypers, 1968).
Electrical stimulation of these cortical motor areas produces predominantly contralateral movements (Asanuma, 1973; Asanuma and Rosen, 1972; Boudrias and others, 2010b;
Kwan and others, 1978; Penfield, 1954; Penfield and Boldrey, 1937; Welker and others,
1957; Woolsey and others, 1952). Although the majority of corticospinal fibers originating from the motor cortex terminate in the contralateral spinal cord, about 5-10% of all corticospinal fibers terminate ipsilaterally (Brosamle and Schwab, 1997). These ipsilateral projections are particularly prevalent from SMA, where 23% of the corticospinal projections are ipsilateral, whereas only 10-13% of corticospinal projections from M1 are ipsilateral (Dum and Strick, 1996; Lacroix and others, 2004; Rosenzweig and others, 2009; Yoshino-Saito and others, 2010). Studies examining muscle activity in both upper limbs (ULs) following SMA stimulation have demonstrated both ipsilateral and bilateral UL movements in addition to purely contralateral movement (Brinkman and
71
Porter, 1979; Mitz and Wise, 1987; Tanji and others, 1988). Recording studies also report that neural activity in the premotor areas is more likely to be associated with ipsilateral or bilateral UL movement compared to activity from M1 (Kermadi and others,
1998; Tanji and others, 1988).
Many functional UL tasks engage a variety of proximal and distal muscles in both
ULs in bilateral movement patterns. Yet most of the literature on control of movement from cortical motor areas has focused on how activity from one or two of the main cortical motor areas influences isolated distal movements in the contralateral UL
(Asanuma and Rosen, 1972; Hendrix and others, 2009; Maier and others, 2002a;
Rizzolatti and others, 1988; Sato and Tanji, 1989). In those studies that have recorded electromyographic (EMG) responses in muscles following stimulation of the ipsilateral corticospinal tract or motor cortex, only muscle activity in the distal UL has been recorded (Aizawa and others, 1990; Soteropoulos and others, 2011). No study to date has specifically compared the motor outputs of these three cortical motor areas for the recruitment of muscles, including proximal and distal muscles, in both ULs. This knowledge is important not only because such movements are common during everyday tasks, but also because these ipsilateral and bilateral outputs may have important implications for mechanisms of recovery from stroke. Further, motor outputs from SMA and PMd appear to have a relatively strong in influence on proximal UL musculature
(Kurata and Tanji, 1986; Macpherson and others, 1982; Tanji and Kurata, 1979). Thus, a more complete understanding of the functions of these motor areas requires inclusion of proximal UL muscles in the analysis.
72
The purpose of the present study was to determine the relative contributions of three cortical motor areas, M1, SMA, and PMd, to recruitment of muscles throughout both ULs, including muscles acting from the shoulder girdle to the wrist. Our hypothesis was that sampling responses in this set of muscles, with responses measured by EMG, would reveal stronger contributions to ipsilateral control than previously recognized from the cortical motor areas. Portions of these results have been reported previously
(Montgomery and others, 2010).
Materials and Methods
Subjects
Three male monkeys (M. fascicularis) were subjects for this study. All subject care complied with the NIH Guide for the Care and Use of Laboratory Animals and all protocols were approved by the Institutional Animal Care and Use Committee at The
Ohio State University. Subjects performed an instructed-delay bilateral reaching task as detailed previously (Davidson and Buford, 2006) designed to provide detectible changes in muscle activation from baseline EMG. Briefly, the task involved the subject reaching for a target on a computer screen located in front of him once a cue was given.
Stimulation was given at any time during the task as long as the subject was in the process of moving his UL.
Surgeries to implant chronic EMG and recording chambers were performed while subjects were under isoflurane anesthesia (1-2%). Vital signs were monitored throughout
73 the procedure. A craniotomy was made over the left cortical motor areas centered near the precentral dimple (around AP 15 ML -12). A 37 x 37 mm square plastic recording chamber (Alpha Omega, Alpharetta, GA) was placed at a 20° lateral tilt over the craniotomy. Another craniotomy was made over the right occipito-parietal region and a recording chamber was placed over this allowing for access to the ponto-medullary reticular formation, for data collection not related to this study. Cranial implants and
EMG connectors were embedded in dental acrylic. Chronic EMG implants made from pairs of Teflon coated stainless steel wires (38 gauge, CoonerWire, Chatsworth, CA) were implanted into 24 selected muscles (listed below). Following surgery, subjects were given analgesics (buprenorphine and ibuprofen) and antibiotics (chlorofenicol or baytril) for several days to prevent postoperative pain and infection.
Stimulation Techniques
Two to four glass coated tungsten microelectrodes (Alpha Omega, Alpharetta,
GA) were simultaneously positioned in the cortex, with electrode sites concentrated in the shoulder/elbow regions for each cortical motor area. Once background neural activity was detected, this was taken as a relative reference point for the penetration depth. Cells not related to UL movements were abandoned and other cells sought. Stimulation was applied to sites where neurons with UL-related activity during task performance were found.
Stimulation threshold for evoking movement was determined in response to a 36 pulse stimulus train (biphasic, 200 us per phase, 333 Hz). The objective was to enable
74 direct comparison of responses from the three motor areas with an identical stimulation paradigm. This pulse train duration was chosen for effective stimulation in premotor areas (Mitz and Wise, 1987).
Stimulation thresholds were defined as the lowest current (± 5 μA) that produced a small but visible muscle twitch. The subject was performing the task as thresholds were determined. Stimulation was delivered using a Master-8 stimulator (A.M.P.I,
Jerusalem, Israel) and a current controlled stimulus isolator (AM-Systems model 2200,
Carlsborg, WA). Thresholds varied between cortical motor areas with most currents used in M1 being between 20-40 μA, those in SMA being between 50-120 µA, and those in
PMd being between 30-80 μA; these current levels are consistent with previous reports
(Hummelsheim and others, 1986; Weinrich and Wise, 1982). After the stimulation threshold for each electrode was found, a series of 10-12 stimulus trains was delivered and EMG responses were recorded.
EMG
EMG was recorded using a Power 1401 CED data acquisition system (CED,
Cambridge, UK). EMGs were obtained and analyzed from flexor carpi ulnaris (FCU), extensor carpi radialis (ECR), biceps brachii (Bic), triceps brachii (Tri), middle deltoid
(MDelt), supraspinatus (supra), upper trapezius (UpTrap), cervical paraspinals
(CervPara), pectoralis major (Pec), and latissimus dorsi (Lats). Subjects H and O had
EMG leads in sternocleidomastoid (SCM) and subjects N and O had EMG leads in lumbar paraspinals (LumPara). In subject N, EMG from left and right teres major were
75 recorded and, in subject H, EMG from left and right brachioradialis were recorded.
However, because data was only collected from one animal for each of these muscles, data from these muscles was eliminated from the analysis.
Tests for EMG integrity
Every two weeks, the EMG leads were tested to confirm they were still in the correct muscles. If the threshold current was > 2000 µA or the muscle response was not in the muscle initially implanted, then the data were omitted from analysis for time points following this observation. In O, left Lats and left CervPara were removed from analysis; in H right CervPara was removed from analysis. Because the removal of CervPara from these subjects led to an imbalance in the chances for left and right CervPara to produce effects, all data for CervPara for both subjects was eliminated from the analysis. EMG implants in H for left FCU and ECR failed near the end of the study, so this data was removed from analysis for the last three sessions. Left and right LumPara were removed from data analysis for N for the last two sessions.
To test for EMG crosstalk, we used an approach described by Cheney and Fetz
(Cheney and Fetz, 1985). A customized script was written using Spike 2 software (CED,
Cambridge, UK). Representative files were chosen from the first and last few weeks of the experiment. EMG waveforms resembling single motor unit action potentials were located for each muscle and these were then used as triggers for the averaging of activity in all muscles studied. If the waveform from the muscle used as the trigger was evident in another muscle, and that waveform had an amplitude greater than 15% of that in the
76 triggering muscle, this was considered crosstalk. There was only one case that met this test; in subject H, right MDelt was removed from analysis due to cross talk with right supra.
Data Analysis
Each EMG waveform was a rectified average of the evoked activity that occurred during the 10-12 stimulus trains applied through each cortical electrode over a 300 msec window. An 89 msec period, before stimulus onset, was used to calculate the EMG baseline mean and standard deviation (s.d.) for all muscles, for each file. The baseline period ended 10 msec before stimulus onset, to avoid any effect of stimulus artifact.
Facilitation was detected by a conventional approach, defining facilitation as the period where the EMG level was elevated by at least 2 s.d. above baseline, and where the increase reached a peak of at least 4 s.d above the mean.
Suppression was more difficult to detect because, in many cases, there was relatively little muscle activity at the time of stimulation. In examining the EMG records visually, we found many instances that appeared to be a clear case of suppression, but which the conventional 2 s.d. criteria could not detect. In these cases, we observed that although the mean EMG level decreased only slightly, the variability of the EMG level was markedly reduced. In order to objectively measure this so as to detect these instances of apparent suppression, we developed a new approach.
EMG data were processed to replot the data as the s.d. of the EMG level taken over a moving 5 msec window. This allowed us to measure variations in EMG level,
77 which added an additional parameter for the detection of suppression. Two of the authors
(LRM, JAB) performed sensitivity and specificity analyses of all suppression responses to compare the visual inspection with a variety of objective criteria. We repeated the sensitivity and specificity analyses for all facilitation responses to compare the conventional objective criteria to our visual inspection.
To detect suppression, the criteria we chose were a combination of two factors.
First, the variability of the EMG level in response to stimulation had to decrease by 1.6 s.d below the variability observed during baseline (this is comparable to the threshold for a one-tailed t-test). Second, the actual EMG level during suppression had to be less than the mean EMG level. The specificity for detection of suppression responses overall was
0.997 (ranging among the subjects from 0.995 to 0.999). The sensitivity for suppression responses was 0.517 (ranging from 0.444 to 0.571). In comparison, the specificity for facilitation responses was 0.993 (ranging from 0.988 to 0.997) and the sensitivity was
0.919 (ranging from 0.891 to 0.948). The sensitivity for suppression could have been improved by making our acceptance criteria more lenient. However, this would have lowered the specificity for suppression, and this was deemed unacceptable. With the criteria chosen, the likelihood of false positive responses for facilitation and suppression was comparable, and both were less than 1%. Figure 4.1 illustrates our approach to suppression detection. Figure 1a illustrates a suppression response which was not detected using the conventional approach but was easily detected by measuring the decreased variability of the EMG level (Fig. 1b) in combination with a slight reduction in
EMG amplitude. There were also times when visual inspection was at odds with the
78 computer response detection (Fig. 1C-F). Although the computer detection method did not always agree with visual inspection, these occasions were rare and involved small suppression responses. Thus, the computer detection approach provided an objective measure of events that worked for suppression responses that were readily apparent on inspection but not detected with the conventional ± 2 SD approach.
Figure 4.1 Detecting suppression events. The graphs show the time that the stimulus train started and finished as well as the baseline mean and the z score of -1.6. A, C, and E represent the average EMG waveforms. B, D, and F represent the average standard deviation of the EMG responses for a 5 msec moving window. In A, the quieter period during the train appears to indicate suppression, but the waveform does not stay below the -1.6 s.d. cut off line, and thus would not be considered a suppression response using a conventional analysis method. In B however, the suppression event is well below the - 1.6 s.d. level. Suppression not only reduces the amplitude of EMG, it also reduces its variability. C shows a response which was detected by visual inspection, however the computer analysis (D) did not detect the event so it was not included in the analysis. In E, no event was detected on visual inspection, yet computer analysis (F) detected the event so it was included in the analysis. 79
To assess the latency of EMG response onset, a customized script was written using Spike 2 software which allowed us to identify the time when the response left baseline and when it returned to baseline.
Anatomical Reconstruction
After euthanasia and perfusion, the brain tissue was cut coronally on a freezing microtome at 50 μm. Every tenth section was mounted and stained with cresyl violet, for reconstruction of the stimulation sites. Tissue sections were compared to the Szabo and
Cowan atlas (Szabo and Cowan, 1984) to locate each motor area. In O, four sites were removed from analysis because stimulation was in subcortical white matter. In addition, seven sites were removed from O as they crossed the midline during penetrations targeting SMA, and had most likely ended up in SMA on the other side.
Statistical analysis
Statistical analyses were run using Microsoft Excel and SPSS version 19 software packages. Chi-square analysis was used to identify differences between frequencies of responses by stimulation site or by laterality. Kruskal-Wallis analysis was used to compare the onset latencies of EMG responses between all cortical areas, and Mann-
Whitney tests were used to compare to median threshold currents and EMG onset latency between individual cortical areas. Descriptive statistics were used to further explore differences in response patterns between cortical regions. Criterion level of p ≤ 0.05 was
80 used to determine statistical significance and post-hoc bonferroni corrections were made to account for multiple comparisons within each group.
Results
Responses from each area
Among the three subjects, a total of 269 sites were stimulated in the three cortical motor areas (M1, SMA, and PMd), with 109 sites from O, 57 sites from N, and 103 sites from H. The majority of sites were in M1 (110 sites), with 88 sites being in PMd and 71 sites located in SMA. Of these, stimulation of 156 sites produced at least one significant response in EMG. Accepted muscle responses were found from 59 of the 110 M1 sites,
50 of the 88 PMd sites, and 47 of the 71 SMA sites. The location of these responsive sites is illustrated in Fig.4.2. Stimulation in each of the three cortical areas resulted in a mixture of ipsilateral, contralateral, and bilateral responses. M1 stimulation evoked the highest number of muscle responses per effective site, averaging 3.9 responses per site, whereas PMd and SMA evoked fewer muscle responses, averaging 2.3 and 2.7 responses per site, respectively. An example of EMG responses from stimulation in SMA is shown in Fig. 4.3, where muscle responses are highlighted by boxes.
By the final acceptance criteria, 473 significant EMG responses were detected.
These came from 58% (156) of the sites tested. Facilitation was the most common response, with 352 of all 473 muscle responses being facilitation (74%) and only 121 responses being suppression (26%). These were distributed evenly throughout the three
81
Figure 4.2 Schematic representation of the cerebral cortex illustrating responsive stimulation sites in each motor area. A is a cranial view of the left cerebral cortex showing the position of M1, SMA, and PMd in relation to physical landmarks (from left to right is caudal to rostral and the lateral surface is uppermost). B shows the caudal to rostral representation of the location of cortical stimulation points that elicited EMG responses (cortical tracings based on Szabo and Cowan). White circles indicate ipsilateral response sites, black triangles indicate contralateral sites, and grey squares indicate bilateral sites. As shown, most of the ipsilateral sites were located in rostral and medial cortical areas (SMA). CS = central sulcus, prcs = precentral dimple, and AS = arcuate sulcus.
82
Figure 4.3 An example of an EMG response from a stimulus train applied in SMA showing EMG responses for 22 muscles. This is a bilateral response with responses seen in both left (L) and right (R) upper limbs. The boxes highlight the objectively detected muscle responses. For example the response in R Lats is a suppression event.
subjects – O (116 facilitation and 40 suppression responses), N (134 facilitation, 39 suppression responses), and H (102 facilitation and 42 suppression responses).
83
M1 and SMA had the highest proportion of facilitation events, with 80% of responses from M1 and SMA being facilitation and 20% being suppression. In contrast,
PMd stimulation caused facilitation in only 58% of responses and suppression in 42% of responses. A chi-square analysis showed there were significantly more EMG suppression events following PMd stimulation compared with stimulation of the other two areas (χ2 =
21.54, p < 0.001). We also compared the proportion of facilitation and suppression events in the ipsilateral and contralateral ULs within each cortical area. In each cortical area the total proportions were similar, with 80% facilitation evoked in the ipsilateral UL and 73% facilitation in the contralateral UL.
As mentioned in the methods, threshold current was used to stimulate each cortical area. As expected, the median current used to stimulate M1 was significantly lower (30 µA) than the median current used in PMd (50 µA, p < 0.001) and SMA (70
µA, p < 0.001). Within each cortical area, there was no difference between threshold currents required to elicit ipsilateral and contralateral responses.
Ipsilateral, bilateral, and contralateral responses by cortical motor area
Sites where stimulation responses were detected were categorized as producing either contralateral (affecting only right UL muscles), ipsilateral (affecting only left UL muscles), or bilateral responses (affecting left and right UL muscles). Only three of the
156 sites (2%) that showed EMG activity were observed to produce bilateral or ipsilateral movement (one site from M1 and two sites from SMA). As illustrated in Fig.4.4, over
84
Figure 4.4 Breakdown of the percentages of muscle responses by laterality for each of the cortical motor areas. M1 stimulation resulted in no ipsilateral responses (A), however stimulation of PMd and SMA (B and C) resulted in a number of ipsilateral responses (C). More than 65% of sites in M1 and PMd produced contralateral responses (A and B), however only 49% of sites in SMA produced contralateral responses (C). The graph in (D) shows the response laterality for each area with the number of sites showing contralateral, ipsilateral, and bilateral responses recorded above each bar. There were more ipsilateral and fewer contralateral sites found in SMA compared with M1 and PMd. For each graph, contralateral responses are represented in black, ipsilateral responses are represented in white, and bilateral responses are represented in grey.
half of the stimulation sites in M1 and PMd produced only contralateral responses (M1: n
= 45, 76% and PMd : n = 34, 68%). Sites producing ipsilateral responses in these areas
85 were not commonly observed, with no M1 sites and 7 PMd sites (14%) producing ipsilateral UL muscle activity. In contrast, SMA sites showed a greater propensity for ipsilateral responses, with 11 SMA sites (23%) producing only ipsilateral activity and 23
SMA sites (49%) producing only contralateral activity. The proportion of bilateral responses evoked from each of the cortical motor areas was similar, with 24% of sites in
M1 (14/59), 18% of sites in PMd (9/50), and 28% of sites in SMA (13/47) evoking EMG responses in both ULs. In all, the proportion of ipsilateral responses resulting from SMA stimulation was higher than the proportion of ipsilateral responses evoked from M1 or
PMd stimulation. Conversely there were significantly fewer contralateral responses evoked from SMA stimulation compared to M1 and PMd stimulation (χ2 = 16.87, p =
0.002).
Although all three areas had a similar proportion of sites producing bilateral responses, we investigated whether there was a difference in the number of ipsilateral responses detected at bilateral sites. M1 and PMd produced a similar ratio (1:3) of ipsilateral to contralateral responses at bilateral sites. The ratio of ipsilateral to contralateral responses at bilateral sites in SMA appeared to be higher (1:2.1), but this difference was not statistically significant.
Ipsilateral v. contralateral responses for individual muscles
The comparison above characterized the laterality of responses evoked from individual stimulation sites. In order to estimate the overall role of the three cortical areas, we next examined the total number and proportion of ipsilateral and contralateral
86
EMG responses in individual muscles from each cortical motor area. Of the 231 responses from M1, 196 (85%) were contralateral and 35 (15%) were ipsilateral. A similar pattern was observed for PMd; 93 of the 114 responses (82%) were contralateral and 21 (18%) were ipsilateral. In SMA, however, a greater proportion of ipsilateral responses was observed; only 84 of the 128 responses (66%) were contralateral while 44
(34%) responses were ipsilateral (χ2 = 18.92, p < 0.001).
We also analyzed the data to determine whether the laterality of evoked responses was related to whether muscles were more proximal or distal within the UL. As shown in
Table 1, the muscle responses were separated by anatomical regions of the trunk and UL:
(1) axial – muscles of the trunk, (2) girdle – muscles around the scapula and shoulder girdle, (3) arm – muscles controlling the shoulder and/or elbow joints, and (4) forearm – muscles in the forearm controlling the wrist. The total muscle responses were evenly distributed between the more proximal (axial and girdle) and more distal (arm and forearm) regions, with 253 of the 473 responses in the axial and girdle regions and 220 responses in the arm and forearm regions. As described previously, of all muscle responses 373 (79%) were contralateral and 100 (21%) were ipsilateral (Fig. 4.5).
Comparing laterality of response by body region revealed that the muscles that were more proximal in the UL (axial and girdle) produced a higher proportion of ipsilateral responses (66/253, 26%) than those that were more distal (34/220, 15%). This difference was significant (χ2 = 24, p < 0.001). Further analysis also revealed there was a difference in the proportion of ipsilateral to contralateral responses in the axial muscles (41/108,
87
Arm Muscle All cortical areas M1 PMd SMA Total Region Ipsi Contra Ipsi Contra Ipsi Contra Ipsi Contra Axial CervPara 6 6 4 4 0 2 2 0 12 LumPara 13 12 2 5 2 0 9 7 25 SCM 4 7 1 6 1 0 2 1 11 Pec 18 42 3 21 6 13 9 8 60 Proportion 38 62 22 78 38 63 58 42 Girdle UpTrap 9 24 4 16 0 6 5 2 33 Supra 12 36 7 19 2 8 3 9 48 Lats 4 60 4 29 0 17 0 14 64 Proportion 17 83 19 81 6 94 24 76 Arm MDelt 7 31 2 14 2 8 3 9 38 Bic 6 24 3 15 0 6 3 3 30 Tri 5 49 0 25 3 15 2 9 54 Proportion 15 85 8 92 15 85 28 72 Forearm FCU 8 41 2 23 1 6 5 12 49 ECR 8 41 2 18 5 13 1 10 49 Proportion 16 84 9 91 24 76 21 79 Total 100 373 34 195 22 94 44 84 473 Table 4.1 List of muscles from which EMG was recorded, grouped according to the body region. The numbers in the ipsilateral and contralateral columns are the total number of EMG responses for each muscle over all three subjects. Muscle responses are shown for all cortical areas combined (shaded grey) as well as for each cortical area independent of the others. The proportion of ipsilateral and contralateral responses is also reported under each muscle region (in italics).
38%) as compared to the proportion of ipsilateral to contralateral responses in other arm regions (59/365, 16%). As shown in Table 4.1, the proportion of ipsilateral responses is higher in the more proximal UL regions for all three cortical areas. This feature is most clearly demonstrated in the responses to M1 stimulation, where only 9% (4/45) of responses in the forearm were ipsilateral compared to 22% (10/46) of responses that were ipsilateral in the axial region.
88
Finally, we tested for differences in the onset latencies of the EMG for contralateral and ipsilateral responses among the motor areas. M1 stimulation produced both the fastest onset times (contralateral responses = 38 +/ - 20 msec) and the slowest onset times (ipsilateral responses = 58 +/- 26 msec). This difference between the contralateral and ipsilateral onset times from M1 stimulation was significant (H = 10.26,
Figure 4.5 Graph showing the breakdown of ipsilateral and contralateral responses based on UL region for all the cortical areas combined (overall) and for each cortical motor area (M1, PMd, and SMA). The numbers above each bar in the graph represent the number of responses detected. The greatest proportion of ipsilateral responses occurred in the most proximal muscle groups (axial region; 38%) compared with more distal areas, including the girdle muscles (17%), arm (15%) and forearm (16%). The hatched line separates the overall responses from those for the individual cortical areas. SMA stimulation resulted in significantly more ipsilateral responses especially in the axial muscle compared to those for M1 and PMd. 89 p = 0.001). There was no difference between the onset of ipsilateral and contralateral responses in PMd (contralateral = 42 +/- 21 msec and ipsilateral = 47 +/- 20 msec) or
SMA (contralateral = 52 +/- 23 msec and ipsilateral = 44 +/- 25 msec). However, there was a significant difference in the onset of contralateral responses from SMA compared to contralateral responses from M1 (W = 24876, p < 0.001) and PMd (W = 7230, p =
0.0021). Overall, contralateral responses from M1 and PMd had the shortest latencies, whereas contralateral responses from SMA and all ipsilateral responses had longer latencies. In terms of the onset latency for facilitation and suppression events, suppression responses elicited by M1 stimulation occurred significantly earlier than facilitation responses (facilitation = 43 msec and suppression = 31 msec; H = 15.81, p <
0.001). This faster suppression from M1 was only significant in the contralateral UL (H
= 14.93, p < 0.001) and was seen in all muscle groups except for muscles of the arm region (axial: H = 4.41, p = 0.036, girdle: H = 5.62, p = 0.018, and forearm: H = 7.13, p =
0.008). As noted above there was no significant difference in facilitation and suppression response onset latency with PMd and SMA stimulation.
Discussion
The present study reveals a somewhat higher degree of ipsilateral and bilateral motor outputs from the cortical motor areas than previously described. Most likely, this is a result of three factors. First, and most obvious, we included both the ipsilateral and contralateral limbs in our study. Second, rather than relying completely upon
90 observation, we used EMG measures to reveal changes in muscle recruitment that might not have been visible as movements in response to stimulation. And third, we sampled muscles throughout the limb, especially proximal muscles, rather than focusing on the distal muscles of the forearm and hand. We also found two specific differences in the distribution of motor outputs from the three cortical motor areas. First, outputs from
SMA were more likely to be ipsilateral and bilateral than those from M1 or PMd. And second, there was slightly more suppression from PMd than from SMA or M1. Finally, this study is consistent with previous reports in the general finding that from all three cortical motor areas, the predominant response observed was contralateral.
The fact that SMA produces more ipsilateral and bilateral outputs extend results of earlier work showing that SMA is a significant source of ipsilateral and bilateral control of the upper limbs (Brinkman and Porter, 1979; Hoshi and Tanji, 2004; Tanji and others, 1988), by demonstrating for the first time that SMA may provide a greater influence over the ipsilateral UL than PMd or M1. Our finding that SMA has a greater influence over the ipsilateral UL when compared to PMd and M1 is important because cortical lesion studies in humans have shown that SMA activity is important in recovery of UL function post injury (Kimberley and others, 2006; Mintzopoulos and others, 2009).
In comparing the three motor areas in the same subject we also found that PMd stimulation resulted in significantly more suppression events compared with stimulation in the other areas. This finding is interesting in regards to previous findings that have shown that some mirror neurons in the ventral premotor area (PMv) result in suppression
91 of pyramidal neurons (Kraskov and others, 2009). This incidental finding of suppression events elicited from PMd stimulation warrants further study.
Physiological studies of the role of the corticospinal tract in animals have traditionally focused on EMG responses in the contralateral limb (Cheney and Fetz,
1985; Fetz and Cheney, 1980), and have rarely included comparisons with ipsilateral
EMG activity. Often, the only record of ipsilateral muscle activity is through visible observation of ipsilateral movement following stimulation (Kwan and others, 1978;
Penfield, 1954; Welker and others, 1957). This method has led to few clear observations of ipsilateral movement and thus resulted in the conclusion that ipsilateral corticospinal activity is relatively inconsequential for motor control (Allison and others, 2000;
Soteropoulos and others, 2011). Although there have been several studies published with careful mapping of the outputs from M1, PMd, and SMA (Asanuma and Rosen, 1972;
Boudrias and others, 2010a; Brinkman and Porter, 1979; Hummelsheim and others, 1986;
Kwan and others, 1978; Rizzolatti and others, 1988), previous studies have not allowed a direct comparison of the effects of identical stimulation trains in all three areas in the same subject. Extrapolating comparisons among studies into a comprehensive picture is difficult. Differences in the degree of neuronal excitability of the M1 and premotor areas are well established, with M1 responding to shorter stimulus trains and lower currents than SMA and PMd (Asanuma and Rosen, 1972; Hummelsheim and others, 1986; Mitz and Wise, 1987; Weinrich and Wise, 1982). The favored neurophysiological approach has been to use the shortest duration stimulus train capable of producing a consistent response (Strick, 2002). However, Asanuma and colleagues and Jankowska et al have
92 shown that repetitive stimulation with longer trains is more effective at producing muscle contraction (Asanuma and others, 1976; Jankowska and others, 1975). Mitz and Wise
(Mitz and Wise, 1987) found that a 36-pulse stimulus train was required to effectively observe the outputs from SMA. Longer trains have also been used in PMd reports
(Hummelsheim and others, 1986; Weinrich and Wise, 1982) and some M1 studies (Kwan and others, 1978). In the present study, we used a 36 pulse (105 ms) train in each cortical motor area in order to allow a direct comparison of the evoked effects among all three areas. At each site, the current used was the minimum amount sufficient to consistently produce a visible motor response. The results obtained must be considered with those methodological details in mind.
Another caveat for the reader is that, as noted in the methods, the prevalence of suppression of muscle activity in response to cortical stimulation was most likely underestimated in the present dataset. We did use the same criteria among all three cortical motor areas, so this does support the conclusion that suppression was more prevalent from PMd than from M1 or SMA. However, our criteria were aimed at specificity, not sensitivity. Comparisons of the overall prevalence of suppression from this study should be interpreted with that in mind.
We note here that, by observations made during the studies, contralateral UL movement was often observed and ipsilateral UL movement was rarely seen. Because we relied upon EMG analysis rather than simple observation, we demonstrate here for the first time that ipsilateral muscles are being recruited even at times when ipsilateral UL movement was not observed. Notably, many of the ipsilateral muscle responses occurred
93 in conjunction with contralateral muscle responses, creating a bilateral pattern that has not been documented previously. In addition, we have demonstrated that the majority of recorded ipsilateral responses occurred in proximal musculature around the trunk and shoulder girdle. Movement in this area is harder to observe than that occurring more distally and may have been overlooked in previous studies that relied on observation of muscle activity.
Our findings, that 21% of responses to stimulation are ipsilateral, seem to contradict earlier physiological studies (Soteropoulos and others, 2011; Tanji and Kurata,
1981; Tanji and others, 1988), which found < 4% of responses were ipsilateral. Our proportion of ipsilateral responses is also higher than would be predicted from tract tracing studies. Only 10-13% of all corticospinal fibers descend ipsilaterally from the motor areas directly to the cervical spinal cord (Lacroix and others, 2004; Rosenzweig and others, 2009; Yoshino-Saito and others, 2010), although the proportion of ipsilateral terminations from SMA is closer to 23 % (Dum and Strick, 1996). We theorize that this is due to our use of longer trains that activate both monosynaptic and polysynaptic pathways (Asanuma and others, 1976; Jankowska and others, 1975; Patton, 1982) and our focus on more proximal muscles of the trunk and UL. Earlier studies have shown that ipsilateral and bilateral activity is more common in proximal muscles than in distal muscles (Bawa and others, 2004). Thus, we would expect a study with a strong representation of proximal muscles to have good success revealing ipsilateral influences.
The control of the trunk and shoulder is imperative in order to allow wrist and hand movements to occur, so including these muscles is important for understanding the role
94 that the ipsilateral cortex may have on the control of UL movement. And as noted, stimulus trains can effectively activate both monosynaptic and polysynaptic pathways
(Asanuma and others, 1976; Jankowska and others, 1975; Patton, 1982). Thus, a number of these ipsilateral responses may have resulted from interhemispheric transcallosal pathways, corticoreticulospinal pathways through brainstem motor nuclei, or pathways at the segmental level via commissural spinal interneurons. Interestingly, studies have shown that although M1 and PMd have transcallosal projections to homologous regions of the opposite cortex, SMA projects to multiple motor areas in the opposite cortex including SMA, PMd, and M1 (Fang and others, 2008; Liu and others, 2002). Such transcallosal projections may explain our findings that SMA produces more ipsilateral and bilateral responses than M1 and PMd, because of the wider effects that SMA has on the contralateral motor cortex. Studies have also shown that the premotor areas also send collateral projections to brainstem nuclei such as the pontomedullary reticular formation
(Keizer and Kuypers, 1984; Keizer and Kuypers, 1989; Matsuyama and Drew, 1997; Rho and others, 1997), and it is well known that commissural interneurons in the spinal cord are often involved in efferent circuits that influence muscle activity (Jankowska and
Edgley, 2006; Jankowska and Stecina, 2007). The fact that higher currents were required to elicit responses from PMd and SMA, compared with the currents required for M1, imply that these premotor areas are more likely to use such indirect pathways involving interhemispheric, brainstem, and/or spinal circuits (Jankowska, 1975).
Although there was no difference in the threshold currents required to elicit ipsilateral or contralateral responses, the onset of ipsilateral responses was significantly
95 later than those of contralateral responses. This delay in response for ipsilateral muscles indicates that the ipsilateral responses were either elicited through direct pathways involving small, lightly myelinated fibers or indirect polysynaptic pathways such as those discussed above. We also found that the latency for suppression was somewhat faster than that for facilitation with M1 stimulation. We suspect that this is due to the general neurophysiological principle that inhibitory synapses tend to be somatic, so inhibition may require less temporal summation and become evident sooner within the train of stimulation. Why this was true for M1 but not PMd or SMA is unclear, but could be an effect of more direct pathways for M1. The design of this study cannot address which of these mechanisms is responsible for our findings. Our hypothesis is that the present results reflect an amalgamation of all of these motor pathways with both direct and indirect pathways being recruited by the stimulus train. It would be interesting to employ retrograde transynaptic tracer injections to muscles which produced the most ipsilateral responses as a strategy to identify these polysynaptic pathways. It would also be interesting to repeat this study comparing responses to short v. long stimulus trains to see whether the prevalence of ipsilateral and bilateral responses was markedly affected.
The current study provides definitive electrophysiological evidence that primary and premotor cortical areas can both facilitate and suppress muscle activity in the ipsilateral and contralateral UL, and identifies clear differences between influences of the three cortical regions. Stimulation of all three regions generates activity in both ULs, but
SMA provides the greatest ipsilateral input to the activity of the muscle groups examined, and PMd activity appears to provide more suppression of EMG than M1 or SMA. These
96 findings demonstrate a functional substrate for ipsilateral cortical control of UL movements in the primate. Further studies of the interactions of these systems may provide a basis to develop strategies for improved motor control of UL movements following injury or damage to the cortex.
97
Chapter 5 : Projections from the supplementary motor area to reticulospinal cells in the
PMRF in a non-human primate as revealed by anterograde and retrograde tract tracing.
Studies investigating the control of voluntary reaching have traditionally focused on connections between the motor cortex and the spinal cord (the corticospinal tract), especially the role the corticospinal tract plays in finger and hand control (Asanuma,
1973; Boudrias and others, 2010b; Cheney and Fetz, 1985; Fetz and Cheney, 1980; He and others, 1993; Kwan and others, 1978; Maier and others, 1993; Ralston and Ralston,
III, 1985; Schieber, 1990; Schieber, 2001). The corticospinal tract, however, is not the only motor system that elicits motor responses in the upper limb (UL). Such alternate motor pathways are involved in postural and gross movements, and arise from nuclei in the brainstem and include the reticulospinal and rubrospinal tracts (Buford and Davidson,
2004; Cheney and others, 1991; Davidson and Buford, 2006; Drew and Rossignol,
1990a; Fujito and Aoki, 1995; Fujito and others, 1991; Hongo and others, 1965; Kuchler and others, 2002; Schepens and Drew, 2004; Van Kan and McCurdy, 2002). Because the motor cortex plays such a dominant role in distal UL function, these brainstem pathways are often overlooked when studying upper limb motor control. In addition, because these alternate pathways do not originate from cells in the motor cortex it is often assumed that movements produced by these pathways are not involved in volitional motor control.
98
There is, however, evidence in the literature that these alternate pathways do receive motor cortical input, and are well positioned to contribute to the execution of volitional movement (Humphrey and Rietz, 1976; Kably and Drew, 1998; Keizer and Kuypers,
1984; Keizer and Kuypers, 1989; King and others, 1972; Matsuyama and Drew, 1997;
Monakow and others, 1979; Newman and others, 1989; Rho and others, 1997).
In particular, several studies in various animal models have provided compelling evidence that the pontomedullary reticular formation (PMRF) receives input from the motor areas of the cerebral cortex (Kably and Drew, 1998; Keizer and Kuypers, 1984;
Keizer and Kuypers, 1989; Matsuyama and Drew, 1997; Newman and others, 1989; Rho and others, 1997). The PMRF, containing cells that are the source of the reticulospinal tract, receives input from four prominent motor areas – primary motor cortex (M1), supplementary motor area (SMA), dorsal premotor area (PMd), and the cingulate motor area (Keizer and Kuypers, 1984; Keizer and Kuypers, 1989). Such projections form the corticoreticular pathway along with projections from other parts of the frontal lobe. Of these cortical motor areas, studies have shown that projections from the motor cortex to the PMRF arise mostly from premotor areas (SMA, PMd, and CMA) compared to M1
(Keizer and Kuypers, 1984; Keizer and Kuypers, 1989; Matsuyama and Drew, 1997; Rho and others, 1997).
Unfortunately none of these studies have detailed the nature of the relationship between projections from the motor cortex and reticulospinal cells. These previous studies have used either retrograde tracers (to label corticoreticular cell bodies in the cortex) or anterograde tracers (to label corticoreticular fibers in the PMRF). No study to
99 date has used a combination of anterograde and retrograde tracers to identify corticoreticular fibers and reticulospinal cells in the PMRF. The combined use of anterograde and retrograde tracers is important for detailing the corticoreticular system as the PMRF consists of interneurons (such as burst cells) as well as reticulospinal cells
(Hikosaka and Kawakami, 1976; Hikosaka and Kawakami, 1977; Van Gisbergen and others, 1981). In light of this, a differentiation between reticulospinal cells and other neurons must be established in order to accurately detail whether corticoreticular projections are coming close enough to reticulospinal cells to make a functional connections or whether their actions are limited to influencing interneurons within the
PMRF. In this study we used retrograde tracers to identify reticulospinal cells in the
PMRF and anterograde tracers to identify corticoreticular fibers originating from SMA so that we can study the organization of corticoreticular projections that directly influence reticulospinal cells.
Although earlier studies demonstrating the presence of a corticoreticular pathway showed that this pathway arose predominantly from premotor areas, these studies were carried out in animal models in which the differentiation of areas within the premotor cortex are difficult. Only one of these studies was performed in a non-human primate model (Keizer and Kuypers, 1989), where there is clear delineation between each of the motor areas allowing for detailed investigation of the contribution from each area to the
PMRF. This study, by Keizer and Kuypers, provided strong evidence that SMA in particular is a source of a number of corticoreticular projections. There are a number of similarities in SMA and reticulospinal function that indicate that these two areas may
100 interact to produce motor outputs. Firstly, although SMA is a motor area, it often shows pronounced activity during preparation for movement – times when postural readiness is required to maintain postural control during movement (Hoshi and Tanji, 2004; Tanji and
Shima, 1994; Tanji and others, 1996). The reticulospinal tract has also been shown to be involved in preparatory postural movements (Buford and Davidson, 2004; Schepens and
Drew, 2003; Schepens and Drew, 2004; Schepens and Drew, 2006). Secondly, physiological evidence indicates that SMA is known to produce not only contralateral limb movement, but also a small but significant amount of ipsilateral and bilateral movement (Brinkman and Porter, 1979; Hoshi and Tanji, 2004; Tanji and others, 1988).
Reticulospinal activity also elicits bilateral muscle activity in the upper limbs (Davidson and Buford, 2006; Herbert and others, 2010; Peterson and others, 1979; Schepens and
Drew, 2006). Thirdly, SMA is involved in producing muscle activity in predominantly proximal muscles (Hummelsheim and others, 1986). This attribute is also seen in the reticulospinal tract, where muscle responses resulting from reticulospinal activity are predominantly in proximal muscles around the upper trunk and shoulder girdle in the
ULs (Buford and Davidson, 2004; Davidson and Buford, 2006; Herbert and others,
2010).
Based on the fact that significant projections have been shown to exist between premotor cortical areas and the PMRF, and that outputs from SMA stimulation resemble outputs from PMRF stimulation, we hypothesized that SMA would send direct projections to the PMRF and, in particular, to reticulospinal cells in the non-human primate. We also hypothesized that the larger reticulospinal cells would receive a greater
101 number of direct projections, as these cells are known to be involved in a number of postural responses (Lingenhohl and Friauf, 1992; Yeomans and Frankland, 1995).
Materials and Methods
Two subjects adult females (M. fascicularis) aged 3-5 years old were used in this study. All subject care complied with the NIH Guide for the Care and Use of Laboratory
Animals and all protocols were approved by the Institutional Animal Care and Use
Committee at The Ohio State University.
Cortical Surgery
To label corticoreticular projections from SMA to the PMRF, subjects were injected with an anterograde tracer – 10% biotinylated dextran amine (BDA) or 10% fluoroscein dextran amine (FD). Subjects received an injection of ketamine (5 mg/kg,
IM) and were intubated and placed on isoflurane anesthesia (1-2%). An injection of dextramethasone (0.7 ml IM) was given preoperatively to prevent cerebral edema. They were then placed in a stereotaxic frame (Kopf instruments, Tujunga, CA). Craniotomies were centered over the left and right frontal cortex at stereotaxic coordinates AP 18 and
ML 11 allowing access to the left and right SMA. Ronguers were used to open up the craniotomies rostrally and caudally so the entire extent of SMA was accessible and the dura was removed. Cortical landmarks (precentral dimple and arcuate sulcus) were used as the lateral and anterior boundaries for SMA.
102
Tracer injections were only made in regions of SMA related to UL movement, thus intracortical microstimulation (ICMS) was first applied to SMA to identify those sites that elicited UL movement. During ICMS, subjects were placed on an IV ketamine
(12 mg/kg/hr) and dexdomedetomidine (1-2 mcg/kg/hr) anaesthesia regime. ICMS was performed using a single tungsten electrode (FHC Corporation. Bowdoin, ME) and electrode placement was based on pre-operative MRI images for depth and the cytoarchitectonic location of SMA as detailed in published primate atlases (Paxinos and others, 2000; Szabo and Cowan, 1984). Electrode tracts were separated by 1 mm in the
AP and ML directions. 3-4 tracks were run at each AP location starting 2.5 mm from midline. To access both the mesial and lateral surface of SMA, electrode penetrations and tracer injections were made at a 20˚ angle. Following the stimulation map of Mitz and Wise (Mitz and Wise, 1987) we created a similar map for each subject, such that sites eliciting arm responses were bound rostrally by at least 2 mm of face responses or no responses, and caudally with 2 mm of leg/tail responses or no responses. Electrode penetrations ranged from AP 22 rostrally to AP 12 caudally. Once the anaesthetic plane was stable ICMS was initiated. The electrode was advanced to multiple depths at each track and a 36 pulse stimulus train (330 Hz, 0.2 msec biphasic pulse) was applied at each depth starting at 2.5 mm from the cortical surface. Currents that elicited observable movement ranged from 50-200 μA.
Following ICMS, subjects were returned to isoflurane (1-2%). Tracer injections of 10k BDA (Life technologies, Carlsbad, CA) 10% in sterile phosphate buffer (PB) were made in the arm regions of SMA in one hemisphere, and injections of a mixture 3k and
103
10k FD (Life technologies, Carlsbad, CA) 10% in sterile PB were made in the arm regions of SMA in the other hemisphere. This combination of tracers has been successfully used before in the non-human primate (McNeal and others, 2010; Morecraft and others, 2002; Morecraft and others, 2007). Tracers were injected using 10 μL microsyringes with a 28 gauge needle (Hamilton, Reno, NV). The syringe was placed at the correct stereotaxic location according to the stimulation map created, and the needle was advanced to the correct depth where the arm movement was elicited. Once the needle had been in place for one minute, 0.4 μL of tracer was then pressure injected into the tissue. The needle was then left for two minutes at the injection location and then moved to the next injection site. In all, subject A had 14 SMA sites injected with FD and
18 SMA sites injected with BDA. Subject G had 19 sites injected with FD and 16 sites injected with BDA. Once all arm sites were injected with either BDA or FD, the craniotomies were covered with gel film (Pharmacia and Upjohn Co, Kalamazoo, MI) and a cap made of dental acrylic (Kooliner, GC America Inc., Alsip, IL). Subjects were then recovered and received a postoperative course of antibiotics (florfenicol – 0.45 ml
SQ) and analgesics (buprenorphine – 0.15 ml IM and metacam – 0.2 ml PO) for 4-7 days.
Spinal Cord Surgery
Four weeks following the SMA tracer injections, subjects then received a laminectomy at C5-6 for retrograde tracer injections to label reticulospinal cells. Details of the laminectomy surgery have been described previously (chapter 3). Briefly, subjects were intubated, placed on isoflurane (1-2%) and placed in a stereotaxic frame. A
104 laminectomy was performed at C5-6 and the spinal cord was exposed. Unilateral injections of 1% cholera toxin subunit B (CTb) (Sigma Aldrich, St. Louis, MO) or 3% fluorogold (FG) (Fluorochrome, Denver, CO) were made in the spinal cord using a 10µL microsyringe and 28 gauge needle (Hamilton, Reno, NV). Following the tracer injections, subjects recovered and given the same course of analgesia and antibiotics as they received following the cortical surgery.
Tissue Processing
Two weeks following the spinal cord surgery subjects were sedated with
Nembutal (3 ml IV) and then transcardially perfused with 0.1M phosphate buffered saline
(PBS), followed by 4% periodate-lysine-paraformaldehyde (PLP), and finally a 10% and
30% sucrose solution. The brain and cervical cord were removed and placed in a solution of 4% PLP overnight. The tissue was placed in a solution of 30% sucrose and left refrigerated. The tissue was then frozen according to the protocol detailed by Rosene and colleagues (Rosene and others, 1986). A freezing microtome was used to section the tissue at 50 µM. Tissue was collected in a series of ten sections. One series of sections was stained with cresyl violet. The other series were used for histochemistry and immunohistochemistry.
Histochemistry and Immunohistochemistry
Tracer injection sites
105
Details of staining for spinal cord injection sites has been previously described
(chapter 3). For the cortical injection sites a series of sections were stained for BDA and
FD. Sections were first rinsed in PBS and incubated in 30% hydrogen peroxide and methanol for 15 minutes. Following more rinses the tissue was left for an hour in a 5% normal horse serum (NHS) blocker (5% NHS/1% bovine serum albumin (BSA)/1% triton
X-100 in PBS). Sections were then incubated in 1:1000 mouse anti-fluorescein primary antibody (EMD Millipore, Billerica, MA) for one hour at room temperature (RT) and then overnight at 4°C. Sections were rinsed in PBS and incubated in avidin-biotin complex (ABC) (Vector labs, Burlingame, CA) for 1 hour. BDA was visualized by placing sections in 3,3’-diaminobenzidine with nickel enhancement (DAB-Ni) (Vector labs, Burlingame, CA) for five minutes. Following several more rinses, sections were incubated for 90 minutes in biotinylated horse anti mouse secondary antibody (1:600;
Vector labs, Burlingame, CA) and placed in ABC for an hour. Sections were then placed in Vector Very Intense Purple (VIP) (Vector labs, Burlingame, CA) for two and a half minutes to visualize the FD. Sections were then mounted and coverslipped. An example of the BDA and FD injection sites for each subject is shown in figure 5.1. Due to the size of the cortical sections, images of cortical injection sites were taken using a Nikon digital camera (Nikon Corporation, Melville, NY).
Brainstem sections
A series of brainstem sections from the caudal pons to the caudal medulla were stained for BDA, FD and either CTb (for subject G) or FG (for subject A). Sections were
106
Figure 5.1 Photomicrographs showing anterograde injection sites in the left and right SMA for subject A and subject G. Both injection sites remained within SMA both the more surface (lateral) layers and the mesial (inner) layers as well as some of the fibers in the more medial white matter. There is a halo surrounding the injection sites that does extend out further. As mentioned, the number of injection sites on each side were similar, however the sites with BDA (in black on the right in each photomicrograph) seemed to contain much more tracer than those with FD (in purple on the left in each photomicrograph). At the level of the brainstem this level of staining intensity was also apparent. After discussions with Dr Robert Morecraft, who has used these tracers extensively in the same animal model, it was suggested that the post operative antibiotic regimen used may have affected the FD. In light of this fact, FD was removed from the analysis. Scale bar is 1 mm.
rinsed and placed in 30% hydrogen peroxide and methanol for 15 minutes. After further rinses, sections were placed in a Bloxall blocking solution (Vector labs, Burlingame, CA) for 30 minutes. Sections were rinsed and placed for an hour in a 5% NHS blocker (as described above), followed by incubation in goat anti-CTb (1:40,000; List Biological
Labs, Campbell, CA) for one hour at RT and then at 4°C overnight, or incubation in rabbit anti-FG (1:10,000; Fluorochrome, Denver, CO) overnight at RT. After several rinses, sections were left in ABC for an hour and then DAB was used to visualize BDA fibers. The tissue was rinsed and incubated in a biotinylated horse anti-goat secondary antibody (1:600; Vector labs, Burlingame, CA) for 90 minutes in subject G or
107 biotinylated horse anti-rabbit secondary antibody (1:600; Vector labs, Burlingame, CA) for 90 minutes for subject A. After several rinses, the tissue was placed in an ABC alkaline phosphatase (ABC-AP) solution (Vector labs, Burlingame, CA) for one hour.
Vector blue was the chromagen used to visualize CTb or FG, so tissues were placed in a solution of vector blue and levamisole (Vector Labs, Burlingame, CA) for 20 minutes.
Sections were rinsed and an avidin- biotin blocker solution (Vector Labs, Burlingame,
CA) was used to block further reactivity prior to incubation of the tissue in a mouse anti- fluorescein primary antibody (1:1000; EMD Millipore, Billerica, MA) overnight at 4° C.
The tissue was rinsed and incubated in a biotinylated horse anti mouse secondary (1:600;
Vector labs, Burlingame, CA) for 90 minutes before being placed in ABC for an hour.
FD was visualized by placing the sections in V-VIP for 2 ½ minutes. Sections were then mounted and coverslipped.
One series of sections was stained for substance P and serotonin (5-HT) to identify various nuclei and landmarks within the brainstem which would then serve as boundaries for the PMRF nuclei. Details of the staining protocol for substance P and 5-
HT have been previously described (chapter 3).
Reticulospinal cells have been noted to have long and complex dendrites
(Scheibel and others, 1973; Scheibel and others, 1955; Valverde, 1961). Thus, some sections were stained for microtubule associated protein 2 (MAP-2) along with BDA and
FG to determine an acceptable distance between corticoreticular terminal swellings and reticulospinal cells that would most likely represent a contact with the soma or its dendrites at least 80% of the time. Sections were stained for BDA and FG as described
108 above with the exception that BDA was visualized with DAB-Ni and FG was visualized with DAB. Sections were then incubated in a mouse anti-MAP-2 primary antibody
(1:40,000; Sigma Aldrich, St Louis, MO) for one hour at RT and then overnight at 4°C.
Sections were rinsed and incubated in a biotinylated horse anti mouse antibody (1:600;
Vector labs, Burlingame, CA) for 90 minutes followed by ABC for 60 minutes. MAP-2 was visualized by placing sections in V-VIP for 2 ½ minutes. Sections were then mounted and coverslipped.
Immunofluorescence
To provide further evidence that corticoreticular swellings that were within close proximity of reticulospinal cells were forming probable contacts with these cells, several sections were stained for synaptophysin (SYNP) in addition to staining for BDA and FG.
To improve the penetration of the antibody through the 50 µM sections, sections were preincubated for seven days in 0.1M ethylenediaminetetraacetic acid (EDTA; Fisher
Scientific, Waltham, MA) prior to staining (Morgan and others, 1997; Morgan and others, 1994). To allow for colocalization of BDA and SYNP, immunofluorescence was used to label all three targets. As FG has a broad emission spectrum which may interfere with the spectrums of the other two labels, it was decided to treat the sections with a
0.5% cupric sulfate solution in ammonium acetate buffer (pH 5.0) for 90 minutes to eliminate the FG fluorescence (Schnell and others, 1999; Tsai and others, 2001).
Sections were then rinsed and placed in a hydrogen peroxide/PBS solution for 15 minutes. After several rinses the tissue was left in a normal donkey serum (NDS; EMD
109
Millipore, Billerica, MA) blocker (5% NDS/1% BSA/1% triton X-100 in PBS) for one hour. Tissue was incubated overnight at RT in rabbit anti-FG primary antibody (1:5000,
Fluorochrome, Denver, CO). Following several rinses, sections were incubated in streptavidin alexa fluor 633 (1:150; Life technologies, Carlsbad, CA) for two hours to label BDA fibers. Sections were then rinsed and left in a solution of donkey anti-rabbit conjugated with alexa fluor 488 (1:200; Life technologies, Carlsbad, CA) for two hours to label cells containing FG. Sections were incubated for 48 hours in a mouse anti- synaptophysin antibody solution (1:5000; EMD Millipore, Billerica, MA). Following several rinses, the tissue was finally incubated in a solution containing donkey anti- mouse conjugated to alexa fluor 555 antibody (1:200; Life technologies, Carlsbad, CA) for two hours. Sections were then mounted and coverslipped with immunomount (Fisher
Scientific, Waltham, MA).
Data Analysis
Stereological approach
An unbiased stereological approach was used in order to estimate the number of labeled corticoreticular swellings located in the PMRF from the ipsilateral and contralateral SMA. The same optical fractionator procedure detailed earlier (chapter 3) was also used for this data set. Briefly each tenth section was stained for BDA, FD, and either CTb or FG. The region of interest (ROI) of the PMRF was then selected using boundaries identified in the 5-HT/substance P stained sections and two macaque brain atlases (Paxinos and others, 2000; Szabo and Cowan, 1984). The boundaries that were
110 used to determine each nucleus within the PMRF are also detailed in chapter 3. To determine the best size counting frame we used the technique outlined by West (West,
1993). We calculated the coefficient of error (CE) for each subject to determine the precision of our estimates according to the formula outlined by West (West and others,
1991). For subject G the CE was 0.03 using 13 sections for analysis, and for subject A the CE was 0.04 using 17 sections for analysis. The mean section thickness was 30 μM.
Quantitative counts of BDA, FD, and CTb or FG labeled objects were made using brightfield microscopy and the Stereoinvestigator system (Stereoinvestigator version 10;
MBF bioscience, Williston, VT). Each section examined was separated from the next by
0.5 mm. Using a Nikon eclipse E800 microscope and a SPOT RTse CCD camera
(Diagnostic Instruments Inc., Sterling Heights, MI), the ROI was outlined for each section. The optical fractionator method was used to estimate the total number of cells in each ROI for each section. Using the stereoinvestigator program, 250 µM x 250 µM counting frames were randomly placed within the ROI for each section within a 750 μM x 750 μM sampling grid. A 5% grid zone was made at the top and bottom of the section.
Figure 2 shows the random allocation of counting frames over all sections within an entire PMRF nucleus for one subject. As illustrated, the frames were distributed throughout the nucleus. All counting was performed at 60x magnification. Different markers were used to identify labeled reticulospinal cells, swellings with BDA that were
≤ 5 μM from labeled cells, swellings with FD that were ≤ 5 μM from labeled cells, swellings with BDA that were > 5 μM from labeled cells, and swellings with FD that were > 5 μM from labeled cells. A swelling was defined as an object labeled with BDA
111 or FD that was at least twice the size of a labeled axon. Most swellings were en passant, although a few were terminal swellings. Line measurements were made between swellings that came close to cells and the cell soma in order to ascertain if they were ≤ 5
μM from the cell. Reticulospinal cell area was measured using the nucleator application from stereoinvestigator. Cell diameter was measured across the largest part of the cell soma.
Determining probable contacts
The only way to determine if SMA projections were making direct synaptic contact with reticulospinal cells would have been to perform electron microscopy, however the nature of this study and the questions posed in this study rendered this approach impractical. Thus other approaches were taken to provide evidence that projections from SMA most likely make contact with reticulospinal cells. The first approach used was to label some sections with BDA, FG, and MAP-2. As the retrograde tracers used to label reticulospinal cells only fill the soma and proximal dendrites, staining with MAP-2 increased the visualization of the dendrites allowing us to calculate an acceptable distance for BDA swellings to be from labeled reticulospinal soma and still be over a reticulospinal dendrite. Fifty labeled reticulospinal cells were chosen throughout these sections. Using light microscopy at 60x magnification, the distance from BDA swellings (that were close to the soma) to the soma was measured. Swellings were either designated as (1) appeared to touch the reticulospinal soma or a MAP-2 stained dendrite of the cell, or (2) not touching the soma or dendrite. For each section,
112 the number of swellings that appeared to touch and those that didn’t touch were then calculated and divided by the total number of swelling counted for that section. If the ratio of the number of swellings touching was less than 0.8 (meaning that less than 80% of the time these swellings appeared to make contact) then the swellings that were furthest away were eliminated until this ratio reached the 0.8 threshold. At 15 µM from the cell, the ratio of the number swellings appearing to make contact was 0.76 (160 swellings contacting and 50 swellings not contacting). At 10 µM the ratio was 0.78 (136 swellings contacting and 39 swellings not contacting). It was not until we reached 5 µM that the ratio reached an acceptable level (0.9) with 98 swellings appearing to make contact and 11 not contacting. Thus, at a distance of ≤ 5 µM between the BDA swelling and the soma we were confident that 90% of the time these BDA swellings would be making contact with the reticulospinal soma or a dendrite from that cell. As a result of this analysis, BDA swellings that were ≤ 5 µM from a reticulospinal cell were classified as probable contacts (contacts).
Our second approach was to stain a series of sections with BDA, FG or CTb, and
SYNP. A number of studies have used synaptophysin or other similar presynaptic proteins such as synapsin, to identify whether projecting fibers are forming functional connections but without quantification of this staining (Deng and others, 2013; Lee and others, 2013; Rosenzweig and others, 2010). SYNP is a protein found in presynaptic terminals and has been used in numerous studies as indirect evidence of synaptic activity between fibers and neurons (Calhoun and others, 1996; Mokin and Keifer, 2006;
Rosenzweig and others, 2010; van den Brand and others, 2012). Once stained, z stack
113 confocal images were taken from random sections using an Olympus FV 1000-spectral confocal system (Olympus America Inc., Melville, NY) at 100x magnification. We were able to identify examples of individual SYNP+ swellings that appeared to contact reticulospinal cells with this approach. Unfortunately, quantification of colocalized swellings containing BDA and SYNP was not able to be performed due to photobleaching of the tissue at the high magnifications which were required to clearly visualize SYNP.
Comparing BDA and FD labeling
When analysis of the amount of BDA and FD labeling was performed in the brainstem sections, it was discovered that the amount of FD labeling was significantly reduced compared to BDA labeling even though the number of tracer injection sites were comparable. Following personal communication with Dr Robert Morecraft who has extensive experience using these tracers in the non-human primate, it was discovered that the use of new generation broad spectrum antibiotics may have affected the transport of this tracer. As a result, the data from the FD injections was removed from analysis.
Despite our attempts to randomize the laterality of the anterograde tracer injections and the retrograde injections the BDA injection always ended up ipsilateral to the retrograde injection which transported successfully. Thus, this paper is focused on the ipsilateral corticoreticulospinal pathway (projections from ipsilateral SMA to the ipsilateral spinal cord via the PMRF). The dual anterograde labeling approach provided evidence that
114 some reticulospinal cells are contacted from both sides of the cerebral cortex, but we could not accurately quantify this.
Post microscopy analysis
For each subject, the ROI and markers for each section were saved and imported into Corel Draw X6 (Corel Corporation, Ottawa, ON, Canada). The ROI and markers for all sections within each individual PMRF nucleus were overlaid onto each other and a single nuclear ROI was created by tracing around all sections to form one contour for the
Figure 5.2 Schematic representation of how the ROI contour for each section for each PMRF nucleus was overlaid onto each other. In this case the contours from four sections were overlaid with the centers aligned. From here an overall contour was made that encompassed all four section ROI’s for the ipsilateral PMRF and then another for the contralateral PMRF so that the eventual schematic for this nucleus contained just one contour for the ipsilateral PMRF and one for the contralateral PMRF instead of four separate contours for each. The blue contours represent the ipsilateral PMRF outlines and the red contours represent the contralateral PMRF outlines for each section. The counting frames are also seen with the green and red outlines that were used for stereological counting. The orange contours within the sections represent reticulospinal cell outlines and the grey represents individual BDA swellings. The distribution of the 250 x 250 μM counting frames throughout the four sections of the nucleus eventually led to a representative sampling of the entire nucleus once the four sections were combined together. 115 ipsilateral and contralateral regions for each nucleus. Figure 5.2 shows how the section
ROI’s for one nucleus were overlaid. So that we could elucidate if there was a difference in the distribution of corticoreticular swellings throughout each nucleus, the ipsilateral and contralateral ROI for each PMRF nucleus was divided into dorsal/ventral zones and medial/intermediate/lateral zones by a 2 x 3 grid. This grid allowed each nucleus to be divided into six zones – dorso-lateral, dorso-intermediate, dorso-medial, ventro-lateral, ventro-intermediate, and ventro-lateral zones. This configuration then allowed us to count number of labeled corticoreticular swellings or reticulospinal cells in each region of the nucleus to elucidate if corticoreticular swellings were located in the same regions where reticulospinal cells are located.
Although the grid used to divide each nucleus into these PMRF zones was symmetrical, the contours of the PMRF nuclei were not. Additionally, the randomized placement of counting frames by the software meant that at times only a portion of the frame was in the region of interest. Additionally, we wanted to make sure that the difference in the number of swellings in a given region was not due to a situation where there were more counting frames in a given region (even though the counting frames were randomly assigned and had an equal chance of being placed in any region of the nucleus). In appreciation of these facts, the clustering of swellings within the dorsal and ventral areas was calculated to assess whether the concentration of swellings was different between within different areas of each PMRF nucleus. To calculate this we found the number of swellings in each counting frame and divided this number by the area of the counting frame where they were located. Only the area of the counting frame
116 that was within the region of interest was calculated. Then the median concentration of the swellings for all the counting frames in each of the six zones of each PMRF nuclei was calculated.
Statistical analysis
All data was entered into Microsoft Excel 2007. Minitab version 16 (Minitab inc,
State College, PA) was used for all statistical calculations. Kruskal-Wallis analyses were used to identify differences in the distribution of swellings based on laterality and the rostro-caudal extent of the PMRF. Mann Whitney tests were performed to compare cell distribution within and between individual PMRF nuclei. Single linear regression analysis was used to identify relationships between the distribution of BDA swellings and labeled reticulospinal cells and a spearman correlation analysis was performed to identify if there was a relationship between cell size and the number of probable contacts it received. A significance level of p = 0.05 was set for all tests and Tukey’s tests were performed to account for multiple comparisons.
Results
Projections from SMA to the PMRF
BDA labeled axons were found bilaterally throughout all four PMRF nuclei studied (figure 3). In both subjects there were more BDA labeled swellings in the PMRF nuclei ipsilateral to the BDA injection compared to contralateral. Subject A had a
117 median 70300 swellings per section in the ipsilateral PMRF compared to 58550 contralaterally. For subject G, there was a median of 80750 swellings per section in the ipsilateral PMRF compared to 45700 contralaterally. Although for each nucleus there were more BDA swellings in the ipsilateral PMRF than the contralateral PMRF, this difference in the laterality of BDA swellings was significant only in PnC for subject A
(W = 40.0, p = 0.012) and in rGi for subject G (W = 26.0, p = 0.03). When analyzing the number of BDA swellings in each nucleus, significant differences existed between the different nuclei in both subjects (subject A: H = 14.44, p = 0.002; subject G: H = 16.41, p
= 0.001). As illustrated in figure 5.3, there were significantly more BDA swellings in
PnC and rGi compared to the number of swellings in cPnO (subject A - PnC: W = 53.0, p
= 0.046; rGi: W = 51.0, p = 0.03; and subject G - PnC: W = 25.0, p = 0.012; rGi: W =
21.0, p = 0.002) and in cGi (subject A - PnC: W = 126.0, p = 0.007; rGi: W = 134.0, p <
0.001; and subject G - PnC: W = 79.0, p = 0.017; rGi: W = 84.0, p = 0.002).
Based on the landmarks used to define each PMRF nucleus, the total area for each nucleus differed, with PnC and rGi having a larger area than cPnO and cGi. The area for each nucleus is given in table 1. Due to this difference, we decided to look at the clustering of swellings in each nucleus to determine if the greater number of swellings in
PnC and rGi for each PMRF nucleus was due just to differences in the nucleus size or whether it reflected an increased concentration of swellings in these nuclei in particular.
For both subjects, the density of swellings was greater in PnC (median = 2843 swellings per mm2) and rGi (median = 2704 swellings per mm2) than it was in cPnO (median =
1434 swellings per mm2) and cGi (median = 1669 swellings per mm2). This difference
118
Figure 5.3 Graph showing the number of BDA swellings in each PMRF nucleus for each subject. Swellings were numerous and found in bilaterally in both subjects. There were significantly more swellings found in PnC compared to cPnO (subject A: p = 0.046, subject G: p = 0.012) and cGi (subject A: p = 0.007, subject G: p = 0.017). There were also significantly more swellings found in rGi compared to cPnO (subject A: p = 0.03, subject G: p < 0.001) and cGi (subject A: p < 0.001, subject G: p = 0.002).
in density was significant (subject A: H = 19.99, p < 0.001; subject G: H = 9.97, p =
0.019).
Location of swellings within PMRF nuclei
119
The distribution of BDA swellings throughout the dorso-ventral and medio-lateral extent of each PMRF nucleus was also analyzed to see if there was a pattern in the distribution of SMA projections to the different PMRF nuclei. The ipsilateral and contralateral PMRF for each nucleus was divided into six zones (as described in methods) in the dorso-ventral and medio-lateral directions. The number of BDA swellings in each zone was subsequently calculated. Schematic diagrams showing the distribution of BDA swellings for each PMRF nucleus and each subject are shown in figures 5.4 and 5.5. The grid dividing each nucleus (figure 5.4A and 5.5A) into six is also shown in these figures to illustrate how the swellings were distributed within each of the zones in the nucleus. The BDA swelling estimates for each zone for each subject is also presented in Table 5.1.
With respect to the dorso-ventral distribution of swellings, there was no significant difference in the number of BDA swellings in the dorsal areas of the PMRF nuclei and the ventral areas. This was true for both the ipsilateral and contralateral
PMRF. When the clustering of swellings was assessed however, a difference was noted.
In the ipsilateral PMRF, despite the similar number of swellings in the ventral and dorsal portions of each PMRF nuclei, there was a difference in the clustering of swellings
(subject A: H = 14.98, p < 0.001; subject G: H = 15.50, p < 0.001). For both subjects, the clustering of swellings was greater in the ventral portion of the nucleus (subject A: median 3008 swellings per mm2; subject G: median 2720 swellings per mm2) compared to the dorsal portion (subject A: median = 1976 swellings per mm2; subject G: median =
120
Ipsilateral PMRF Contralateral PMRF Subject A cPnO PnC rGi cGi cPnO PnC rGi cGi Total area (mm2) 10.249 14.041 13.467 8.214 10.249 14.041 13.467 8.214 Swellings per zone DL 28500 14600 11100 4100 16700 16600 25700 3800 DI 50800 82700 77300 37100 48400 77000 78500 16800 DM 31900 70500 88200 27800 31000 78500 77300 20500 VM 37900 116300 70000 15500 10400 54100 44200 25000 VI 56600 143700 103400 26900 7100 64000 82400 46300 VL 18900 80400 50500 8800 1800 39100 40000 35100 Subject G cPnO PnC rGi cGi cPnO PnC rGi cGi Total area (mm2) 7.235 14.647 13.422 6.128 7.235 14.647 13.422 6.128 Swellings per zone DL 18800 40000 11900 1300 8000 9000 9500 6500 DI 31700 81000 65200 15500 12700 55600 52100 15800 DM 4700 40000 89800 20800 11000 46600 41300 17100 VM 12800 87400 119100 18900 15300 62800 50400 17300 VI 32800 177200 97100 45300 25900 59400 79300 22000 VL 21700 66400 27800 20300 8000 25300 26500 11100 Table 5.1 Estimated number of BDA swellings in each zone of each PMRF nucleus for both subjects. Also included in the table is the total area in mm2 for each nucleus for each subject. As illustrated, the number of swellings in the medial and intermediate zones of each nucleus were greater than the number of swellings in the lateral zones. This was true for both the ipsilateral and contralateral PMRF. Abbreviations for each zone are as follows – DL = dorso-lateral, DI = dorso-intermediate, DM = dorso-medial, VM = ventro-medial, VI = ventro-intermediate, and VL = ventro-lateral.
1888 swellings per mm2). Although this difference was seen in all nuclei, it was only
significant in PnC (subject A: W = 171.5, p = 0.011; subject G: W = 485.5, p = 0.002)
and rGi (subject A: W = 521.5, p = 0.004; subject G: W = 593.5, p = 0.036). In the
contralateral PMRF the picture was more complex with subject A having a greater
concentration of swellings dorsally in cPnO (W = 251.0, p = 0.002) and PnC (W = 321.0,
121 p = 0.02), and a greater concentration of swellings ventrally in cGi (W = 112.0, p = 0.03).
However, there were no significant differences between the concentration of swellings dorsally and ventrally in subject G.
In terms of the medio-lateral distribution of BDA swellings, there was a pattern to the distribution of swellings with a greater number of swellings being located in the medial (subject A: median = 8800; subject G: median = 5900) and intermediate zones
(subject A: median = 8700; subject G: median = 7650) of the PMRF nuclei compared to the lateral zone (subject A: median = 4650; subject G = 4900). This difference was statistically significant in both subjects (subject A: H = 9.38, p = 0.009; subject G: H =
11.78, p = 0.003). In the ipsilateral PMRF, when each nucleus was analyzed this difference in medio-lateral distribution was only significant in PnC for subject A (W =
83.0, p= 0.04) and rGi for subject G (W = 87.5, p = 0.002), although there was a trend for the medial and intermediate zones to have a greater number of swellings in rGi for subject A (W = 180.5, p = 0.099) and in PnC for subject G (W = 198.5, p = 0.108). As with the dorso-ventral distribution, we also looked at the density of swellings in each medio-lateral zone for each PMRF nucleus and found that in both subjects the density of
BDA swellings was greatest in the medial (subject A: 2832 swellings per mm2; subject G:
2730 swellings per mm2) and intermediate zones (subject A: 2704 swellings per mm2; subject G: 2480 swellings per mm2) compared to the lateral zone (subject A: 1.891 swellings per mm2; subject G: 1.664 swellings per mm2). However, when analyzed by individual PMRF nuclei, the density of swellings in the medial and lateral zones of rGi was significantly greater only in subject G (W = 86.0, p = 0.0014). In subject A,
122 however, this increase in swelling density was trending towards significance in rGi (W =
160.0, p = 0.063).
In the contralateral PMRF, there was a similar distribution of swellings, with most swellings located in the medial and intermediate zones compared to the lateral zone. This difference was only significant in subject G (H: 8.23, p = 0.016), although in subject A there was a strong trend towards this same distribution (H: 5.61, p = 0.06). When the distribution of swellings was analyzed by PMRF nucleus, this pattern was seen for all nuclei in both subjects although it only reached significance in cPnO in subject G (W =
28.0, p = 0.011). As with the ipsilateral PMRF, we next analyzed the clustering of swellings to see if there was a concentration of swellings in any of these zones. In both subjects there was a higher concentration of BDA swellings in the medial (subject A: median = 1871.5 swellings per mm2; subject G: median = 1992 swellings per mm2) and intermediate zones (subject A: median = 2048 swellings per mm2; subject G: median =
1127 swellings per mm2). This increased concentration of swellings in the medial and intermediate zones compared to the lateral zone was significant in both subjects (subject
A – H: 11.36, p = 0.003; subject G – H: 14.79, p = 0.001). Analyzing the concentration of swellings in each nucleus, the difference between the medial and intermediate zones and the lateral zones was only significant in rGi (subject A – W = 79.0, p = 0.037; subject
G – W = 52, p = 0.004).
Relationship between SMA projections and reticulospinal cells
123
Early reports regarding the morphology of reticulospinal cells based on golgi impregnation noted the extensive network of dendrites associated with these cells
(Scheibel and others, 1973; Scheibel and Scheibel, 1961; Valverde, 1961). These authors often noted that dendrites could be found to extend > 800 µM from the soma and form a web-like network with other reticulospinal cells nearby (Scheibel and others, 1973;
Valverde, 1961). As the retrograde tracers used only fill the soma and some primary dendrites, we are unable to identify if these BDA swellings are making contacts with distal dendrites. However, by detailing if there is a relationship between where the BDA swellings are located and where the reticulospinal cells are located we can provide indirect evidence that projections from SMA may be preferentially located within this reticulospinal network. The same division of each PMRF nucleus into six zones was used to investigate whether the distribution of BDA swellings followed a similar pattern to the distribution of reticulospinal cells. Figures 5.4 and 5.5 also show the distribution of reticulospinal cells throughout each PMRF nucleus along with the distribution of BDA swellings for each subject.
In the ipsilateral PMRF, there was a positive correlation between the number of
BDA swellings and the number of cells in a given zone of the PMRF (subject A – r2 =
0.756, p < 0.001; subject G – r2 = 0.796, p < 0.001). As illustrated in figure 5.6A, for each subject as the number of BDA boutons increases in a given PMRF zone so does the number of reticulospinal cells in the same area. The correlations for the number of BDA swellings and number of reticulospinal cells in each nucleus are illustrated in the graphs
124
125
Figure 5.4 Schematic illustrations of the distribution of BDA swellings and reticulospinal cells throughout the four PMRF nuclei for subject A. In A, the six section grid for the ipsilateral and contralateral PMRF is shown illustrating the six different zones and where they are located in the ipsilateral and contralateral PMRF – dorso-lateral (DL), dorso-intermediate (DI), dorso-medial (DM), ventro-medial (VM), ventro-intermediate (VI), and ventro-lateral (VL). In B, the location of the PMRF nuclei within the brainstem is shown on representative brainstem sections for each nuclei. The distribution of BDA swellings (in grey) and reticulospinal cells (in orange) are shown for each nucleus in C, with the six section grid overlaid to show where each of the regions are. Sampling was made throughout the ipsilateral and contralateral PMRF. There were a greater number of swellings and cells in the medial and intermediate regions of each nucleus compared to the number of swellings and cells in the lateral regions
125
126
Figure 5.5 Schematic illustrations of the distribution of BDA swellings and reticulospinal cells throughout the four PMRF nuclei for subject G. As in figure 4, A illustrates the six section grid for the ipsilateral and contralateral PMRF with the six different zones and where they are located in the ipsilateral and contralateral PMRF – dorso-lateral (DL), dorso-intermediate (DI), dorso- medial (DM), ventro-medial (VM), ventro-intermediate (VI), and ventro-lateral (VL). In B, the location of the PMRF nuclei within the brainstem is shown on representative brainstem sections for each nuclei. The distribution of BDA swellings (in grey) and reticulospinal cells (in orange) are shown for each nucleus in C, with the six section grid overlaid to show where each of the regions are. Sampling was made throughout the ipsilateral and contralateral PMRF. There were a greater number of swellings and cells in the medial and intermediate regions of each nucleus compared to the number of swellings and cells in the lateral regions.
126 in figure 5.6B. Studying each PMRF nucleus separately, this relationship is found in both subjects in PnC (subject A - r2 = 0.951, p = 0.003; subject G – r2 = 0.969, p =
0.001), and cGi (subject A - r2 = 0.897, p = 0.015; subject G – r2 = 0.949, p = 0.004).
The picture for cPnO and rGi was mixed with subject A demonstrating a strong positive correlation in rGi (r2 = 0.822, p = 0.045) and subject G demonstrating it in cPnO (r2 =
0.922, p = 0.009). Although subject G did not show a strong correlation in rGi, there was a weak correlation between the number of BDA swellings and reticulospinal cells (r2 =
0.481). For subject A, in cPnO no correlation was found (r2 = 0.174). In the contralateral PMRF, a weak correlation was found overall in subject G (r2 = 0.626) which, when analyzed by individual PMRF nuclei showed no significance. There was no correlation between the number of swellings and the number of reticulospinal cells in subject A in the contralateral PMRF (r2 = 0.086).
Evidence for contacts between SMA projections and reticulospinal cells
As described in the methods, BDA swellings that were ≤ 5µM from labeled reticulospinal soma were characterized as BDA contacts as there was a 90% chance that they were making direct contact with the soma or dendrites of that cell. Several sections were also stained for BDA, FG, and synaptophysin to provide further evidence that these swellings within 5 μM contained presynaptic proteins indicating that they are most likely forming synaptic contacts with the labeled reticulospinal cell. An example of the colocalization of
127
Figure 5.6 Relationship between location of BDA swellings and location of reticulospinal cells in the ipsilateral PMRF. Each dot represents one of the six zones found in the grid that was show in figs 4 and 5. Thus there are 24 points on each graph in A (6 zones from all four nuclei) and 6 points on each graph in B (6 zones in each nucleus). In A, the overall relationship between swellings and cells is shown demonstrating that throughout the PMRF there is positive correlation between the number of swellings and the number of reticulospinal cells. In B, the correlation between swellings and cells is shown by individual PMRF nucleus for each subject. When analyzed by nucleus this positive correlation between swellings and cells was only significant in PnC and cGi.
128
BDA and synaptophysin in a swelling within 5 μM of a labeled reticulospinal cell is shown in Fig 5.7A.
Using the criteria of a BDA swelling ≤ 5 μM from a reticulospinal cell being a probable contact, a small but noticeable number of BDA swellings formed probable contacts with reticulospinal cells. There were more BDA contacts in the ipsilateral
PMRF compared to the contralateral PMRF, however this is most likely due to the small number of labeled reticulospinal cells in the contralateral PMRF. Over 80% of all BDA contacts in each PMRF nucleus were in the ipsilateral PMRF. Figure 5.7B illustrates the distribution of BDA contacts by PMRF nucleus for each subject. There were significantly more probable contacts in the ipsilateral PMRF compared to the contralateral PMRF in both PnC and rGi (subject A – p < 0.05, subject G – p - <0.05).
As shown in figure 5.7C-D, the proportion of BDA contacts in the ipsilateral PMRF increases in rostrocaudal direction. Only 1-2% of all BDA swellings in the ipsilateral cPnO make probable contacts with reticulospinal cells, yet 6-7% of all BDA swellings in the ipsilateral cGi make probable contacts. This trend was also found in the contralateral
PMRF with < 1% of swellings in cPnO and PnC making probable contact and 2% of swellings in rGi and cGi making probable contact. Figure 5.8 shows that the distances from probable contacts to reticulospinal soma were distributed relatively evenly from 0.1
– 5 μM. The one exception to this was that for all nuclei the highest proportion of probable contacts was at a distance between 1.1 and 2 μM from the soma (fig. 5.8).
129
Figure 5.7 Some BDA swellings come within 5μM of labeled reticulospinal cells making it 90% likely that they are contacting the cell soma or proximal dendrites. The photomicrograph in A shows the colocalization of SYNP (red) with BDA (blue) in a corticoreticular swelling that is < 5 μM from a FG labeled reticulospinal cell (green). The colocalization of SYNP and BDA in this swelling indicates that this swelling has the functional components of a presynaptic terminal and is most likely forming a synaptic contact with the reticulospinal cell. B is a graph showing the number of probable contacts made between BDA swellings and reticulospinal cells. There are a greater number of contacts in the ipsilateral PMRF compared to the contralateral PMRF especially in PnC and rGi. C and D show the proportion of swellings that come within 5μM of a reticulospinal cell (in white) compared with the proportion of swellings that do not (in grey). The proportion of swellings that come within 5μM increases in both subjects as you go from the rostral to the caudal extent of the PMRF.
130
Figure 5.8 Graphs showing distance of swellings from the cell for each PMRF nucleus for each subject (subject A – black, subject G – grey). Swellings were placed in bins based on their distance from the cell (0.1-1μM, 1.1-2 μM, 2.1-3 μM, 3.1-4 μM, and 4.1-5 μM) and the number of swellings in each bin was plotted on the graph. There was a unimodal distribution in all nuclei for both subjects with the greatest number of swellings being 1.1-2 μM from the cell.
131
Figure 5.9 Graphical representation of the number of cells that received probable contacts from corticoreticular swellings. In A, the number of cells with contacts (+) and the number of cells without contacts (-) are shown for each PMRF nucleus for each subject for the ipsilateral and contralateral PMRF. In all nuclei of the ipsilateral PMRF there are a greater number of cells receiving probable contacts than not. This is most marked in PnC and rGi for both subjects. B shows the proportion of cells with probable contacts (grey) and without (white) for each ipsilateral PMRF nucleus for each subject. This shows that for all nuclei there was a higher proportion of cells receiving probable contacts than those that were not.
132
Although there is evidence for probable contacts between SMA projections and reticulospinal cells, further analysis was needed to describe the nature of these projections. Importantly we wanted to know what proportion of reticulospinal cells have probable contacts versus those that do not. In the ipsilateral PMRF, 74-77% of cells received probable contacts (fig.5.9A). From figure 5.9A it is clear that there is more variation in the contralateral PMRF with 57-74 % of cells receiving probable contacts.
This variation is most likely due to the low number of cells in the contralateral PMRF.
Overall, in the ipsilateral PMRF there were significantly more cells with contacts than without for both subjects (subject A - H = 13.16, p < 0.001; subject G – H = 9.28, p =
0.002). Analyzing each nucleus separately, the number of cells with BDA contacts was higher than the number without contacts in all nuclei, although this difference only reached significance in PnC for subject G (W = 10, p = 0.03) and rGi for subject A (W =
15, p = 0.012). However there was a trend towards significance in PnC for subject A (W
= 19.5, p = 0.117) and rGi for subject G (W = 11, p = 0.061). Figure 5.9B illustrates the difference in the proportion of cells with contacts and without contacts for each nucleus and for each subject. In the contralateral PMRF, there was no significant difference between the number of cells with and without boutons for either subject.
Cell size and probable contacts
Given that reticulospinal cells varied greatly in size from small cells (≤ 25 μM) to giant cells (≥ 55 μM), it was of interest to identify if a relationship existed between the
133
Figure 5.10 Graphs illustrating the relationship between the size of the reticulospinal cell (μM) and the number of corticoreticular probable contacts it receives. There was a positive correlation between cell size and the number of probable contacts in the ipsilateral PMRF. In A, the overall relationship between the number of probable contacts and cell size is shown demonstrating that there was a positive correlation between the number of swellings and the cell size. In B, the correlation is shown by individual PMRF nucleus. This positive correlation between swellings and cells was only significant in PnC and rGi. All spearman rho coefficients and p-values are reported for each graph.
134 size of the reticulospinal cell and the number of probable contacts made by SMA projections. Figure 5.10A shows that throughout the ipsilateral PMRF there is a positive correlation between reticulospinal cell diameter and the number of probable contacts for both subjects (subject A - ρ = 0.438, p < 0.001, subject G - ρ = 0.413, p < 0.001). When each PMRF nucleus was analyzed (figure 5.10B), it was discovered that for both subjects this relationship between cell diameter and probable contacts is significant in PnC
(subject A - ρ = 0.518, p < 0.001; subject G - ρ = 0.340, p = 0.01) and rGi (subject A - ρ
= 0.423, p = 0.001, subject G - ρ = 0.473, p < 0.001). In subject A, there was a significant correlation between cell size and the number of contacts in cPnO (ρ = 0.849, p
= 0.008). In subject G, this relationship between cell size and contacts was significant in cGi (ρ = 0.604, p = 0.001). In the contralateral PMRF there was also a positive correlation between cell size and the number of contacts for both subjects (subject A - ρ =
0.534, p = 0.001; subject G - ρ = 0.429, p = 0.025). However, when each nucleus was analyzed individually the only significant relationship between cell size and the number of contacts was found in rGi in subject A (ρ = 0.512, p = 0.001).
Discussion
This manuscript provides the first detailed description of projections from SMA to the reticulospinal system in the PMRF of the non-human primate. Figure 5.11 illustrates the proposed corticoreticulospinal pathway from SMA to the spinal cord via
135
Figure 5.11 Schematic diagram showing the pathway from SMA to the spinal cord via the PMRF. Corticoreticular projections (in red) are bilateral to the PMRF with a slight preference to terminating in the ipsilateral PMRF. Reticulospinal projections (in blue) are bilateral to the spinal cord, although they are predominantly ipsilateral. Thus, although the corticoreticulospinal pathway is a bilateral pathway, there is a preponderance of projections to project ipsilaterally to the spinal cord. In A, the red in the cortex represents SMA and the red lines corticoreticular projections from SMA to the PMRF. The blue lines represent reticulospinal projections from the PMRF to the spinal cord. B represents a flow diagram showing the flow of information from SMA to the PMRF and then to the spinal cord.
136 reticulospinal cells in the PMRF. Corticoreticular projections from SMA to the PMRF are bilateral, although there are more projections to the PMRF ipsilateral to hemisphere from which the projections originated. The projections are found throughout the rostro- caudal extent of the PMRF, however there are significantly more swellings in PnC and rGi. In addition, there is a greater concentration of swellings in PnC and rGi compared to cPnO and cGi. Within the ipsilateral PMRF, SMA projections are more concentrated in the ventral portion of the nuclei, especially in PnC and rGi. SMA projections were also found most commonly in the medial and intermediate regions of the ipsilateral and contralateral PMRF, which is also where a majority of the reticulospinal cells are located.
A small but important proportion of these projections come close enough to reticulospinal cells to form probable contacts with these cells. Most importantly, in all nuclei a majority of the reticulospinal cells that project to the ipsilateral spinal cord receive contacts from SMA fibers. There is also evidence that these probable contacts contain synaptophysin, a presynaptic protein commonly found in presynaptic terminals, providing further evidence that these contacts are most likely forming connections with reticulospinal cells. The proportion of probable contacts increases in a rostral to caudal direction. Finally, there is also a relationship between reticulospinal cell size and the number of probable contacts with larger cells receiving more probable contacts than smaller cells.
Projections from motor cortex to the PMRF
137
A number of studies have provided evidence that projections exist from the motor cortex to the PMRF (Keizer and Kuypers, 1984; Keizer and Kuypers, 1989; Matsuyama and Drew, 1997; Newman and others, 1989; Rho and others, 1997). Although there is evidence of a corticoreticular pathway, the role of this pathway in motor control is still poorly understood. Most of these studies were conducted in rodent and cat models, animals with a less developed motor cortex than humans and in particular a poorly defined premotor cortex. Thus, any conclusions regarding the origin of corticoreticular projections or the strength of this pathway in these animal models has limited translational value to clinical human situations. Although in rodents and cats the premotor areas are poorly developed compared to primates, cortical areas analogous to primate premotor areas (such as the frontal agranular cortex in the rodent and area 6aβ and 6aγ in the cat) have been shown to produce a significant number of corticoreticular projections (Keizer and Kuypers, 1984; Matsuyama and Drew, 1997; Newman and others, 1989; Rho and others, 1997).
The one study that has investigated corticoreticular projections in non-human primates verifies the findings from lower order animals that projections from the premotor cortex are numerous (Keizer and Kuypers 1989). These authors found evidence that the premotor cortex (SMA and PMA) produces more projections to the PMRF than
M1 (Keizer and Kuypers 1989). However, a major caveat of their study was the fact that they only studied labeling in the contralateral cortex. This paper expands on these earlier findings, by demonstrating that there are a significant number of corticoreticular projections from SMA (an area of the premotor cortex) to the PMRF. We have also
138 added further information from this earlier study by showing that although projections from SMA to the PMRF are bilateral, there is a tendancy for the projections to be more ipsilateral. These corticoreticular projections course throughout the rostro-caudal extent of the PMRF from cPnO to cGi. They are also present across the entire dorso-ventral and medio-lateral territories of each PMRF nucleus. Although these corticoreticular projections are found in all PMRF nuclei, there are significantly more swellings found in
PnC and rGi. These two nuclei are also where the majority of reticulospinal cells are located and where the largest reticulospinal cells are also predominantly located
(Lingenhohl and Friauf, 1992; Newman, 1985a; Newman, 1985b; Sakai and others,
2009). Additionally, there is evidence that corticoreticular swellings are located within the same regions of the PMRF nuclei where reticulospinal cells are also found. One may argue that this is mainly due to the quantification approach undertaken (i.e. the cells counted were within the counting frames where the swellings were also counted).
However, our approach of unbiased stereology allowed an equal chance for all areas of the nuclei to be sampled, and as shown in figures 1, 3 and 4 sampling was distributed throughout each of the PMRF nuclei. In addition, it is clear from our quantification that the concentration of the corticoreticular swellings in the counting frames located in the lateral regions of all the PMRF nuclei was much lower than for the intermediate and medial regions. This was also true for the reticulospinal cells with much fewer cells being located in the lateral zones. Therefore we are confident that this increased concentration of corticoreticular swellings in the medial and intermediate zones is a true reflection of the distribution patterns and not due to sampling bias or error.
139
Relationship between SMA corticoreticular projections and reticulospinal cells
Reticulospinal cells are known to have long, complex networks of dendrites and receive synaptic inputs from neurons originating from a number of regions of the nervous system (Eccles and others, 1975; Peterson and others, 1974; Scheibel and others, 1973;
Scheibel and others, 1955; Valverde, 1961). Researchers have noted that reticulospinal dendrites can reach lengths of over 800µM, and early work by Scheibel and colleagues used a conservative estimated length of 100µM for calculations related to reticulospinal dendrites (Scheibel and Scheibel, 1961; Scheibel and others, 1955; Valverde, 1961). The nature of the tracing study undertaken here is that only the soma and most proximal dendrites were labeled, thus our quantification of probable contacts is conservative.
Despite the conservative criteria used to identify corticoreticular swellings that probably contacted labeled reticulospinal cells, there were still a small but significant number of corticoreticular swellings that came within 5µM of reticulospinal cells. The presence of synaptophysin within swellings that are within 5µM of reticulospinal cells provides further support that these swellings are most likely forming a functional connection with the cell. There was also a unimodal distribution of the distance between the corticoreticular probable contacts and reticulospinal cells, with most of the contacts being between 1.1 - 2µM from the cell soma. So, although our criteria for probable contacts extended to swellings that were up to 5 μM from reticulospinal cells most of these contacts came much closer to the cell, within 2 μM in fact.
140
Although there was a fairly even distribution of corticoreticular projections to
PMRF nuclei ipsilateral and contralateral to the hemisphere from where the projections originated, there was a clear preference for corticoreticular contacts to be mostly ipsilateral. This is thought to be most likely due to the fact that there was a clear preponderance for most reticulospinal cells to be found in the ipsilateral PMRF. Perhaps one of the most striking findings in this study is the fact that throughout the PMRF a greater number of reticulospinal cells received SMA contacts than the number that did not. This finding seems to underpin the significance of corticoreticular projections from
SMA to the PMRF and in particular projections from the ipsilateral SMA to reticulospinal cells that project to the ipsilateral spinal cord. If the corticoreticular pathway from SMA to the reticulospinal tract were vestigial or inconsequential in motor control then there would be no reason for a majority of reticulospinal cells to receive this input. This corticoreticulospinal pathway from SMA to the spinal cord via the reticulospinal tract is also significant in that it appears to be a predominantly ipsilateral pathway and thus a way for the ipsilateral motor cortex to influence ipsilateral limb movement. However, such conclusions must be tempered with caution as we were not able to study the contralateral corticoreticular projections to identify the organization of corticoreticular projections to a large number of reticulospinal cells in the contralateral
PMRF. As mentioned in the methods we attempted to answer this question, however our difficulty with the second anterograde tracer (FD) made it impossible to address this question. Further studies using dual anterograde or retrograde tracers are needed in order to clarify whether this ipsilateral dominance in probable contacts is related to a true
141 preference within the corticoreticular system from SMA or whether this is due to the disparity between the number of reticulospinal cells labeled in the ipsilateral and contralateral PMRF.
Our finding that there is a greater concentration of swellings in the same areas of each PMRF nuclei where most of the reticulospinal cells are also found provides indirect evidence that these corticoreticular projections may be making contacts with reticulospinal dendrites further distally, contacts which we have been unable to identify with the criteria that we applied to identify probable contacts. Further investigation is needed to identify if these corticoreticular projections located further distally from the reticulospinal soma are forming probable contacts with the more distal reticulospinal dendrites. In addition, further studies must be undertaken at the electron microscope level in order to identify if these swellings that we have identified as probable contacts are making functional synapses with the reticulospinal cells, or other cells in the PMRF
Analysis also showed that there was a positive correlation between reticulospinal cell size and the number of corticoeticular probable contacts being made with that cell.
At first glance this appears straightforward as the larger cells have a larger surface area, thus there is more room to make contacts and so a greater number of contacts on larger cells would be expected. However, there is not a perfect correlation between cell size and the number of contacts as would be expected if the sole explanation for more contacts on larger cells was related to the larger surface area. Indeed there were some large cells that had few and even no contacts, and alternatively there were some smaller cells that had a large number of contacts. Also the fact that this correlation was only
142 significant in PnC and rGi for both subjects indicates that SMA projections have a preference for large cells in these two nuclei in particular. Concomitantly, PnC and rGi are also the two nuclei which have a significantly greater proportion of cells receiving probable contacts compared to the proportion of cells receiving corticoreticular contacts in cPnO and cGi. We will discuss in the next section the functional significance of these two findings as they relate to the role of SMA and reticulospinal cells in the postural control of the upper limb.
Functional significance of corticoreticulospinal system from SMA
It is well established in the literature that SMA is an area of the motor cortex that is involved in controlling internally guided movements and plays an important role in initiating postural adjustments in anticipation of movement (Hoshi and Tanji, 2004; Tanji and Shima, 1994; Tanji and others, 1996). The reticulospinal tract has also been shown to play a critical role in postural control (Prentice and Drew, 2001; Schepens and Drew,
2003; Schepens and Drew, 2004; Stapley and Drew, 2009). Moreover, activity in SMA and the PMRF preferentially activate proximal musculature around the trunk and shoulder girdle (Buford and Davidson, 2004; Davidson and Buford, 2006; Hummelsheim and others, 1986; Peterson and others, 1979; Peterson and others, 1978). Thus the similarities in function and muscle recruitment patterns from these two distinctly separate motor areas implies that at some level some form of interaction between these two areas occurs. Evidence presented in this study indicates that this interaction is occurring at the
143 level of the PMRF, although this does not exclude the possibility that additional interactions between the two systems are also occurring at the level of the spinal cord.
In order to successfully participate in postural control in anticipation of and during movement, reticulospinal cells receive inputs from multiple regions of the nervous system, from the cerebellum to sensory systems (Eccles and others, 1975; Lingenhohl and Friauf, 1992; Scheibel and Scheibel, 1961; Scheibel and others, 1955; Valverde,
1961; Yeomans and Frankland, 1995; Yeomans and others, 2002). Reticulospinal cells are then tasked with integrating these multiple inputs into a single motor output. Inputs from the motor cortex regarding plans for upcoming movement as well as information regarding motor cortex preparations for movement would be a vital source of information which reticulospinal cells could integrate with inputs it has already received from other neural structures. SMA is an important source for information regarding the preparation for cortically driven movements, thus it is appropriate for strong connections to exist between SMA and reticulospinal cells. That there appears to be a greater number and concentration of SMA projections in PnC and rGi is also significant because these two nuclei are the source of most reticulospinal cells including large and giant cells that participate in postural responses such as the acoustic startle reflex (Beran and Martin,
1971; Lingenhohl and Friauf, 1992; Newman, 1985a; Newman, 1985b; Sakai and others,
2009; Yeomans and Frankland, 1995; Yeomans and others, 2002).
Our finding that larger reticulospinal cells receive more probable contacts reflects a preference for SMA to influence reticulospinal cells that are involved in more dynamic postural adjustments that occur during movements. Cells with large somas have large
144 axons and thus faster conduction velocities (Henneman and others, 1965). Postural adjustments that occur during movements must occur rapidly in response to feedback that the nervous system is receiving during movement. Axons with the fastest conduction velocities are best placed to elicit these rapid postural adjustments. As SMA is plays a significant role in postural responses during movement, a corticoreticulospinal pathway which sends inputs from SMA to the large reticulospinal cells is one way in which SMA can direct postural adjustments during upper limb movements. The smaller cells which also receive SMA inputs may be involved in anticipatory postural adjustments that occur in preparation for movement. These postural adjustments are often more subtle and require the adoption of more sustained muscle contractions and limb positions.
Researchers have shown that small axons in the cortex with their slower conduction velocities are often more active during these slow sustained movements (Evarts, 1965;
Fromm and Evarts, 1981). Thus it is reasonable to suspect that these reticulospinal cells with smaller axons may play a similar role in slow sustained movements involved in postural responses.
In conclusion, the PMRF receives bilateral inputs from SMA and these inputs are concentrated in areas where the reticulospinal cells are also found, namely PnC and rGi.
Some of these SMA projections form probable contacts with reticulospinal cells, predominantly on cells that are ipsilateral to the projections. Although these contacts are found on both small and large reticulospinal cells, there is a preference for more contacts to be located on larger cells. These larger reticulospinal cells are also the cells with the
145 fastest conduction velocity, allowing them to elicit rapid postural adjustments that are needed during upper limb movements to maintain control of the limb.
146
Chapter 6 : Discussion
For motor control of the upper limb, the corticospinal tract, and M1 in particular, has been at the center of most lines of scientific inquiry. This is due primarily to the importance of this pathway in producing purposeful, fine dexterous movements of the hand and fingers (Lemon, 2008b; Maier and others, 1993; Schieber and Santello, 2004).
Although this is clearly an important field of investigation, the focus of upper limb research on the hand may reflect a lack of appreciation for the rest of the upper limb, particularly proximally around the shoulder and trunk, and the role that these structures play in placing the hand close to the object and then providing upper limb stability which allows the intended hand movement to be performed accurately (Castiello and others,
1992; Gentilucci and others, 1992; Huang and Brown, 2013). Motor control around the shoulder and trunk relies on multiple motor pathways including the reticulospinal tract
(Buford and Davidson, 2004; Davidson and Buford, 2006; Fujito and others, 1991;
Herbert and others, 2010; Hummelsheim and others, 1986; Schepens and Drew, 2004;
Wiesendanger and others, 1974). The corticospinal tract is also involved in muscle recruitment around the shoulder and trunk, however premotor cortical areas such as SMA contribute more significantly to the corticospinal control around the shoulder and trunk than they do to the control of the hand (Hummelsheim and others, 1986; Maier and
147 others, 2002b; Wiesendanger and others, 1974). The reticulospinal tract has also been shown to preferentially elicit muscle responses proximally around the shoulder and trunk
(Buford and Davidson, 2004; Davidson and Buford, 2006). Given that both SMA and the reticulospinal tract contribute to muscle recruitment in the proximal upper limb, and that both of these motor systems contribute to postural control (Schepens and Drew, 2003;
Schepens and Drew, 2004; Wiesendanger and others, 1974), the idea that interactions between these two systems may contribute to limb stability during upper limb tasks seems plausible. Despite this apparent connection between SMA and the reticulospinal tract and upper limb stability, this dissertation represents the first body of work to study in detail the connections between SMA and the reticulospinal tract in the non-human primate. The experiments detailed here provide evidence of the existence of an anatomical pathway between SMA, the reticulospinal tract, and the spinal cord. Firstly, a first comprehensive description of the organization of the reticulospinal tract in the NHP is provided. Evidence is also provided that SMA, compared to other cortical motor areas, is most likely to recruit muscles in the ipsilateral upper limb or both upper limbs.
Ipsilateral and bilateral muscle activity is also a feature of reticulospinal tract outputs.
The delayed onset latencies of these ipsilateral muscle responses from SMA imply that these are most likely facilitated through a multisynaptic pathway, such as the one that exists between SMA, the reticulospinal tract, and the spinal cord. Finally, a detailed description of cortical projections from SMA to the PMRF and specifically reticulospinal cells is given.
148
One aspect of the reticulospinal tract within the non-human primate that has remained elusive is the anatomical organization of this pathway. The details of this tract in other animal models has been well documented (Beran and Martin, 1971; Martin and
Dom, 1971; Martin and others, 1979; Matsuyama and others, 1999; Matsuyama and others, 1997; Newman, 1985a; Newman, 1985b) however, the two studies that have described the organization of this pathway in primates have lacked detail and clarity
(Kneisley and others, 1978; Sakai and others, 2009). In particular, an earlier paper from our lab presented unexpected findings regarding reticulospinal organization in the NHP
(Sakai and others, 2009). The data presented in chapter 3 of this dissertation provides some clarity to these earlier results. Our findings show that the reticulospinal tract in the
NHP follows a similar level of organization to the tract in lower order animals. Although the tract is bilateral, there is a strong preponderance cells to project ipsilaterally in the spinal cord. In fact, in PnC and rGi (where most reticulospinal cells are located) > 70% of reticulospinal cells were in the ipsilateral PMRF (ipsilateral to the retrograde injection). Our analysis shows that as Kneisley and colleagues (Kneisley and others,
1978) described, the organization of the reticulospinal tract is phylogentically conserved in the primate.
Although the reticulospinal tract sends bilateral projections to the spinal cord, the majority of the projections descend in the ipsilateral cord. Why is this important?
Physiological studies in primates have shown that muscle recruitment in the upper limbs is bilateral following reticulospinal activity (Buford and Davidson, 2004; Davidson and
Buford, 2006) as it is in other animals (Drew and Rossignol, 1990b). The circuitry
149 involved in producing these patterns of motor output has been studied and detailed in the cat (Bannatyne and others, 2003; Jankowska and others, 2003) and involves a complex network of commissural spinal neurons. This facilitation of bilateral effects via spinal circuitry is also thought to be important in other animal models with a similar anatomical organization in the reticulospinal system as the cat. By showing the organization of the reticulospinal tract in primates parallels that of the cat, we provide evidence that the bilateral reticulospinal effects observed physiologically are most likely also due to spinal circuitry as it is in the cat.
Understanding that these bilateral motor outputs are most likely facilitated through local spinal neurons is particularly important when considering changes that may occur to the reticulospinal tract following neurological injury. Following injury, changes could occur in the reticulospinal tract that may facilitate functional recovery of the upper limb. Such changes would most likely occur at the spinal cord level where axonal sprouting of reticulospinal neurons to innervate other spinal neurons may help facilitate motor outputs to muscles that have lost their primary innervations as a result of the injury. Axonal pruning of reticulospinal neurons may also occur in the spinal cord.
Reticulospinal outputs often show a stereotypical pattern of muscle activation
(Davidson and Buford, 2006; Drew and Rossignol, 1990b) that resembles the flexor synergy patterns that develop following cortical injury such as stroke (Pandian and Arya,
2012). Cortical injuries such as stroke can lead to a loss in the control of inhibition, resulting in recruitment of multiple muscles in concert instead of isolation (Pandian and
Arya, 2012). This leads to poor isolated movement, and this flexor mass pattern of upper
150 limb muscle activity must be overcome in order to regain independent functional control of the affected limb. Changing the patterns of connections between the reticulospinal tract and spinal neurons through pruning could be one way to limit this synergistic pattern that have lost their inhibitory influences from other motor pathways. Even though the data presented here provides indirect evidence that local spinal neurons facilitate the bilateral effects associated with reticulospinal activity, further studies are needed to confirm that this is indeed the case. One study that has already been initiated in our lab as a followup to the initial study performed in chapter 3 is a dual retrograde tracer study using CTb tagged with either Alexa fluor 488 (green) of Alexa fluor 594 (red). The goal of this study is to label reticulospinal cells that project in the left and right spinal cord to identify if some cells send bilateral projections to the spinal cord (these cells would contain both the red and green tracers). Future studies that would follow on from these experiments would involve using of transsynaptic tracers that would be injected into proximal muscles which show the most responses in relation to reticulospinal activity such as supraspinatus and upper trapezius. The aim of this study would be to label the reticulospinal circuitry from the reticulospinal neurons in the PMRF to the motor units in the muscle. Spinal interneurons involved in this circuit would also be labeled. Another interesting question to address would be whether the larger reticulospinal cells labeled with our cervical retrograde tracer injection were involved in postural control. By injecting an anterograde tracer into the dorsal cochlear nucleus (DCN) to label neurons involved in the acoustic startle reflex, and then injecting a retrograde tracer into the cervical cord we could discover if the labeled reticulospinal cells are part of the acoustic
151 startle reflex and thus primed to elicit quick postural movements. Physiologically, intracellular recordings from the spinal cord similar to those carried out in the cat could be undertaken in order to provide direct evidence that such spinal circuitry involving interneurons conveys reticulospinal commands in the primate.
Although the cortical motor areas in the NHP have all been studied to some degree (Lemon, 2008b), a detailed study of M1, SMA, and PMd influences on bilateral proximal upper limb muscle activity in the same animal was lacking. The question of how each of these cortical areas influence proximal upper limb motor control is important for two reasons. Firstly, the reticulospinal system elicits motor responses in mostly proximal muscles of bilateral upper limbs. Thus, studying the influences of three different cortical motor areas on these same proximal upper limb muscles will allow us to those responses with the effects that the reticulospinal tract has on these same muscles.
Secondly, based on previous studies, there is evidence that these three cortical areas send projections to reticulospinal cells (Keizer and Kuypers, 1984; Keizer and Kuypers, 1989;
Matsuyama and Drew, 1997; Newman and others, 1989; Rho and others, 1997), and that the secondary motor areas of SMA and PMd have stronger connections to the PMRF than
M1 (Keizer and Kuypers, 1989). Studying the effects of activity from these three motor areas on proximal muscles of the upper limbs allowed a comparison to determine whether any of these cortical areas have a significant influence of these muscle groups similar to the influence that the reticulospinal tract has.
The data presented in chapter 4 provides evidence that motor outputs from SMA show a similar pattern to motor outputs from the reticulospinal tract in that they activate
152 proximal muscles around the shoulder girdle and trunk in both upper limbs. In particular we found an unexpected 21% of responses were ipsilateral, with a majority of them resulting from SMA activity. The onset latencies for these responses were also somewhat delayed, implying that these responses were facilitated through a polysynaptic pathway such as the corticoreticulospinal pathway. The similarities between SMA and reticulospinal motor outputs on the proximal upper limb, combined with the evidence previously shown that in SMA and the reticulospinal tract have important roles in postural control, lead us to postulate that there may be a common motor pathway shared between SMA and the reticulospinal tract that allows them to perform these shared functions of proximal muscle activity and postural control. It is unclear however, at what level of the nervous system this interaction between SMA and the reticulospinal tract may take place. As stated earlier, there is evidence that SMA sends projections to the
PMRF. However, it is also the case that corticospinal and reticulospinal projections do not often synapse directly on motoneurons, instead synapsing on spinal interneurons
(Bannatyne and others, 2003; Boudrias and others, 2010a; Boudrias and others, 2010b;
Jankowska and others, 2003; Lemon, 2008b; Peterson and others, 1975). For the corticospinal tract, this is particularly true for projections influencing proximal muscles
(Lemon, 2008b; Ralston and Ralston, III, 1985). Thus the interaction between SMA and reticulospinal projections may occur at the level of the spinal cord rather than in the
PMRF. Further studies are needed to elucidate whether a functional connection between
SMA and the reticulospinal tract exists, and whether this interaction occurs in the PMRF or the spinal cord, or both. Currently our lab is analyzing data on combined stimulation
153 in the PMRF and the cortical motor areas (M1, PMd, and SMA) to identify if a combination of cortical and reticulospinal activity can lead to changes in the muscle activation patterns compared to when they are stimulated alone. Another way to study connections between reticulospinal cells and the cortical motor areas is to do some antidromic studies involving PMRF stimulation and recording from cortical areas to see if the interaction between reticulospinal cells and cortical motor areas occurs at the level of the brainstem.
One question that remains from this study is which polysynaptic pathway is involved with producing the ipsilateral motor responses. Although we have hypothesized that a pathway involving the PMRF could be involved with eliciting these responses, another pathway involving transcallosal pathways could also account for these findings
(Fang and others, 2008; Liu and others, 2002). One way to address this is to perform a followup study involving recording from the cortical motor areas in one hemisphere while stimulating in the other. In this paradigm, orthodromic and antidromic activity could be investigated to see if the stimulation applied to cortical areas in one hemisphere are producing activity in the other hemisphere thus indicating that transcallosal projections may be involved in producing the muscle responses seen. By understanding the level of interaction that exists between SMA and the reticulospinal system, and the details about the substrates that allow this interaction to occur, we will be able to start examining how this interaction between these two systems changes following neurological injury and how this level of interaction may be able to be targeted for improved therapeutic interventions.
154
To start to address the question of how this interaction between SMA and the reticulospinal tract may occur, a detailed description of the projections from SMA to the
PMRF and reticulospinal cells is provided in chapter 6. Although previous studies examined cortical projections from the motor cortex to the PMRF, these studies were undertaken in lower order animals such as rats and cats (Keizer and Kuypers, 1984;
Matsuyama and Drew, 1997; Newman and others, 1989; Rho and others, 1997).
Additionally, these studies have not incorporated specific labeling of the reticulospinal cells, meaning that they lack evidence for direct connections between cortical projections and reticulospinal cells. Our study presents the first detailed description of the organization of corticoreticular projections from SMA within the PMRF within the NHP showing that this pathway is bilateral. In this dissertation, we also provide evidence for the first time that about 5-7% of projections from SMA to the PMRF come close enough to form connections with reticulospinal cells and that these projections are mainly ipsilateral. Further, a vast majority of reticulospinal cells in the ipsilateral PMRF (>
73%) had some corticoreticular input.
Although we were able to provide evidence that there were probable contacts between SMA and reticulospinal cells, questions still remain regarding whether these are functional connections. Preliminary evidence suggests that they may be, related to our colocalization of BDA with synaptophysin, however a detailed quantitative analysis of the number of BDA labeled swellings containing synaptophysin would be needed in order to start presenting concrete evidence that such functional connections exist between these two systems. Additionally, the only true way to identify whether there are actual
155 synapses between the corticoreticular projections and reticulospinal cells would to be to perform electron microscopy. Such an experiment would be long and intensive, however it would provide the first direct evidence that these two systems interact with each other.
One followup experiment that would be enlightening would be to perform a triple label experiment to identify the proximal dendrites of the reticulospinal cells using MAP-2.
We did this on a select number of sections to identify a conservative distance that the swellings could be from the cell to still probably contact a dendrite. However, as it is well established that reticulospinal cells have long dendrites (Scheibel and others, 1973;
Valverde, 1961) there may be many probable contacts that we are missing, and these contacts that are more distal on the dendrites may be the site for axonal sprouting following injury. In addition, because of inconsistent labeling with the fluorescein tracer we were never able to study the contribution of projections from the contralateral SMA.
A followup experiment to look at the contralateral projections would be a valuable addition to this body of work.
The fact that we now have evidence of the organization of corticoreticular projections between SMA and reticulospinal cells in the PMRF gives an understanding of where changes may occur following cortical injury such as stroke. This study detailed in chapter 5 describes the organization of the corticoreticular system in the non-injured system, thus allowing us to compare this in the future with the organization of this system following neurological injury. More importantly it provides evidence of a potential pathway that may be important in functional recovery following injury. Many studies in humans have shown that following a cortical stroke affecting M1, changes in the activity
156 of SMA occur (Kimberley and others, 2006; Mintzopoulos and others, 2009; Rehme and others, 2012; Rehme and others, 2011; Riecker and others, 2010). The premise of functional recovery following injury via SMA has always focused on interhemispheric connections between SMA and other cortical motor areas in the opposite hemisphere
(Otsuka and others, 2013; Sung and others, 2013; van Meer and others, 2012; van Meer and others, 2010). However, the evidence provided here opens up a new line of investigation. What if this increased activity in SMA following M1 injury is also leading to increased recruitment of reticulospinal neurons via the corticoreticulospinal pathway?
If this is the case, how could this pathway influence the recruitment of upper limb muscles, especially proximally, to promote upper limb recovery? It is clear that the evidence of the existence of the corticoreticulospinal pathway in the NHP has provided a starting point from which these important clinical questions can be addressed. It is also clear that the potential influence of this pathway on facilitating recovery is significant, and as such it is hoped that future research may start to consider the impact that this system may have on recovery following neurological injury.
The experiments included in this dissertation provide significant insight into the contributions of SMA and the reticulospinal tract to the motor control of the upper limb.
It highlights a level of interaction between these two motor systems that may influence recovery of upper limb function following neurological injury. However, the ways in which this interaction may affect recovery are still unknown. This dissertation provides a foundation that it is hoped will one day form the basis of effective rehabilitative strategies that can be implemented to improve functional outcomes in upper limb function
157 following neurological injury. However, much more study and research is needed in order to reach this ultimate goal.
158
References
Ada L, Goddard E, McCully J, Stavrinos T, Bampton J. 2005. Thirty minutes of positioning reduces the development of shoulder external rotation contracture after stroke: a randomized controlled trial. Arch Phys Med Rehabil 86(2):230-4.
Aizawa H, Mushiake H, Inase M, Tanji J. 1990. An output zone of the monkey primary motor cortex specialized for bilateral hand movement. Exp Brain Res 82(1):219-21.
Allison JD, Meador KJ, Loring DW, Figueroa RE, Wright JC. 2000. Functional MRI cerebral activation and deactivation during finger movement. Neurology 54(1):135-42.
Alstermark B, Isa T, Pettersson LG, Sasaki S. 2007. The C3-C4 propriospinal system in the cat and monkey: a spinal pre-motoneuronal centre for voluntary motor control. Acta Physiol (Oxf) 189(2):123-40.
Asanuma H. 1973. Cerebral cortical control of movement. Physiologist 16(2):143-66.
Asanuma H, Arnold A, Zarzecki P. 1976. Further study on the excitation of pyramidal tract cells by intracortical microstimulation. Exp Brain Res 26(5):443-61.
Asanuma H, Rosen I. 1972. Topographical organization of cortical efferent zones projecting to distal forelimb muscles in the monkey. Exp Brain Res 14(3):243-56.
Baker SN. 2011. The primate reticulospinal tract, hand function and functional recovery. J Physiol 589(Pt 23):5603-12.
Baker SN, Spinks R, Jackson A, Lemon RN. 2001. Synchronization in monkey motor cortex during a precision grip task. I. Task-dependent modulation in single-unit synchrony. J Neurophysiol 85(2):869-85.
Bannatyne BA, Edgley SA, Hammar I, Jankowska E, Maxwell DJ. 2003. Networks of inhibitory and excitatory commissural interneurons mediating crossed reticulospinal actions. Eur J Neurosci 18(8):2273-84.
159
Bawa P, Hamm JD, Dhillon P, Gross PA. 2004. Bilateral responses of upper limb muscles to transcranial magnetic stimulation in human subjects. Exp Brain Res 158(3):385-90.
Beran RL, Martin GF. 1971. Reticulospinal fibers of the opossum, Didelphis virginiana. I. Origin. J Comp Neurol 141(4):453-65.
Bortoff GA, Strick PL. 1993. Corticospinal terminations in two new-world primates: further evidence that corticomotoneuronal connections provide part of the neural substrate for manual dexterity. J Neurosci 13(12):5105-18.
Boudrias MH, Lee SP, Svojanovsky S, Cheney PD. 2010a. Forelimb muscle representations and output properties of motor areas in the mesial wall of rhesus macaques. Cereb Cortex 20(3):704-19.
Boudrias MH, McPherson RL, Frost SB, Cheney PD. 2010b. Output properties and organization of the forelimb representation of motor areas on the lateral aspect of the hemisphere in rhesus macaques. Cereb Cortex 20(1):169-86.
Bouisset S, Zattara M. 1987. Biomechanical study of the programming of anticipatory postural adjustments associated with voluntary movement. J Biomech 20(8):735-42.
Bradnam LV, Stinear CM, Byblow WD. 2013. Ipsilateral motor pathways after stroke: implications for non-invasive brain stimulation. Front Hum Neurosci 7:184.
Brinkman C, Porter R. 1979. Supplementary motor area in the monkey: activity of neurons during performance of a learned motor task. J Neurophysiol 42(3):681-709.
Brinkman J, Kuypers HG. 1973. Cerebral control of contralateral and ipsilateral arm, hand and finger movements in the split-brain rhesus monkey. Brain 96(4):653-74.
Brochier T, Spinks RL, Umilta MA, Lemon RN. 2004. Patterns of muscle activity underlying object-specific grasp by the macaque monkey. J Neurophysiol 92(3):1770-82.
Brodal A. 1956. Anatomical aspects of the reticular formation of the pons and medulla oblongata. Prog Neurobiol(2):240-55.
Broeks JG, Lankhorst GJ, Rumping K, Prevo AJ. 1999. The long-term outcome of arm function after stroke: results of a follow-up study. Disabil Rehabil 21(8):357-64.
Brosamle C, Schwab ME. 1997. Cells of origin, course, and termination patterns of the ventral, uncrossed component of the mature rat corticospinal tract. J Comp Neurol 386(2):293-303.
Buford JA, Davidson AG. 2004. Movement-related and preparatory activity in the reticulospinal system of the monkey. Exp Brain Res 159(3):284-300. 160
Calhoun ME, Jucker M, Martin LJ, Thinakaran G, Price DL, Mouton PR. 1996. Comparative evaluation of synaptophysin-based methods for quantification of synapses. J Neurocytol 25(12):821-8.
Canedo A, Lamas JA. 1993. Pyramidal and corticospinal synaptic effects over reticulospinal neurones in the cat. J Physiol 463:475-89.
Carmel JB, Kimura H, Berrol LJ, Martin JH. 2013. Motor cortex electrical stimulation promotes axon outgrowth to brain stem and spinal targets that control the forelimb impaired by unilateral corticospinal injury. Eur J Neurosci 37(7):1090-102.
Castiello U, Bennett KM, Paulignan Y. 1992. Does the type of prehension influence the kinematics of reaching? Behav Brain Res 50(1-2):7-15.
Cheney PD, Fetz EE. 1985. Comparable patterns of muscle facilitation evoked by individual corticomotoneuronal (CM) cells and by single intracortical microstimuli in primates: evidence for functional groups of CM cells. J Neurophysiol 53(3):786-804.
Cheney PD, Fetz EE, Mewes K. 1991. Neural mechanisms underlying corticospinal and rubrospinal control of limb movements. Prog Brain Res 87:213-52.
Conte WL, Kamishina H, Reep RL. 2009a. Multiple neuroanatomical tract-tracing using fluorescent Alexa Fluor conjugates of cholera toxin subunit B in rats. Nat Protoc 4(8):1157-66.
Conte WL, Kamishina H, Reep RL. 2009b. The efficacy of the fluorescent conjugates of cholera toxin subunit B for multiple retrograde tract tracing in the central nervous system. Brain Struct Funct 213(4-5):367-73.
Cordo PJ, Nashner LM. 1982. Properties of postural adjustments associated with rapid arm movements. J Neurophysiol 47(2):287-302.
Davidson AG, Buford JA. 2006. Bilateral actions of the reticulospinal tract on arm and shoulder muscles in the monkey: stimulus triggered averaging. Exp Brain Res 173(1):25- 39.
Deecke L. 1987. Bereitschaftspotential as an indicator of movement preparation in supplementary motor area and motor cortex. Ciba Found Symp 132:231-50.
Deng LX, Deng P, Ruan Y, Xu ZC, Liu NK, Wen X, Smith GM, Xu XM. 2013. A novel growth-promoting pathway formed by GDNF-overexpressing Schwann cells promotes propriospinal axonal regeneration, synapse formation, and partial recovery of function after spinal cord injury. J Neurosci 33(13):5655-67.
Denny-Brown D. 1966. The cerebral control of movement. Springfield, IL: Charles C Thomas. 161
Dewald JP, Pope PS, Given JD, Buchanan TS, Rymer WZ. 1995. Abnormal muscle coactivation patterns during isometric torque generation at the elbow and shoulder in hemiparetic subjects. Brain 118 ( Pt 2):495-510.
Donaldson C, Tallis R, Miller S, Sunderland A, Lemon R, Pomeroy V. 2009. Effects of conventional physical therapy and functional strength training on upper limb motor recovery after stroke: a randomized phase II study. Neurorehabil Neural Repair 23(4):389-97.
Drew T, Dubuc R, Rossignol S. 1986. Discharge patterns of reticulospinal and other reticular neurons in chronic, unrestrained cats walking on a treadmill. J Neurophysiol 55(2):375-401.
Drew T, Rossignol S. 1990a. Functional organization within the medullary reticular formation of intact unanesthetized cat. I. Movements evoked by microstimulation. J Neurophysiol 64(3):767-81.
Drew T, Rossignol S. 1990b. Functional organization within the medullary reticular formation of intact unanesthetized cat. II. Electromyographic activity evoked by microstimulation. J Neurophysiol 64(3):782-95.
Dum RP, Strick PL. 1991. The origin of corticospinal projections from the premotor areas in the frontal lobe. J Neurosci 11(3):667-89.
Dum RP, Strick PL. 1996. Spinal cord terminations of the medial wall motor areas in macaque monkeys. J Neurosci 16(20):6513-25.
Dum RP, Strick PL. 2013. Transneuronal tracing with neurotropic viruses reveals network macroarchitecture. Curr Opin Neurobiol 23(2):245-9.
Eccles JC, Nicoll RA, Schwarz WF, Taborikova H, Willey TJ. 1975. Reticulospinal neurons with and without monosynaptic inputs from cerebellar nuclei. J Neurophysiol 38(3):513-30.
Ellis MD, Drogos J, Carmona C, Keller T, Dewald JP. 2012. Neck rotation modulates flexion synergy torques, indicating an ipsilateral reticulospinal source for impairment in stroke. J Neurophysiol 108(11):3096-104.
Evarts EV. 1965. RELATION OF DISCHARGE FREQUENCY TO CONDUCTION VELOCITY IN PYRAMIDAL TRACT NEURONS. J Neurophysiol 28:216-28.
Evarts EV. 1968. Relation of pyramidal tract activity to force exerted during voluntary movement. J Neurophysiol 31(1):14-27.
Evarts EV. 1969. Activity of pyramidal tract neurons during postural fixation. J Neurophysiol 32(3):375-85. 162
Fang PC, Stepniewska I, Kaas JH. 2008. Corpus callosum connections of subdivisions of motor and premotor cortex, and frontal eye field in a prosimian primate, Otolemur garnetti. J Comp Neurol 508(4):565-78.
Fetz EE, Cheney PD. 1980. Postspike facilitation of forelimb muscle activity by primate corticomotoneuronal cells. J Neurophysiol 44(4):751-72.
Fetz EE, Cheney PD, German DC. 1976. Corticomotoneuronal connections of precentral cells detected by postspike averages of EMG activity in behaving monkeys. Brain Res 114(3):505-10.
Flash T, Mussa-Ivaldi F. 1990. Human arm stiffness characteristics during the maintenance of posture. Exp Brain Res 82(2):315-26.
Fromm C, Evarts EV. 1981. Relation of size and activity of motor cortex pyramidal tract neurons during skilled movements in the monkey. J Neurosci 1(5):453-60.
Fujito Y, Aoki M. 1995. Monosynaptic rubrospinal projections to distal forelimb motoneurons in the cat. Exp Brain Res 105(2):181-90.
Fujito Y, Imai T, Aoki M. 1991. Monosynaptic excitation of motoneurons innervating forelimb muscles following stimulation of the red nucleus in cats. Neurosci Lett 127(1):137-40.
Gentilucci M, Chieffi S, Scarpa M, Castiello U. 1992. Temporal coupling between transport and grasp components during prehension movements: effects of visual perturbation. Behav Brain Res 47(1):71-82.
Grillner S, Deliagina T, Ekeberg O, el MA, Hill RH, Lansner A, Orlovsky GN, Wallen P. 1995. Neural networks that co-ordinate locomotion and body orientation in lamprey. Trends Neurosci 18(6):270-9.
He SQ, Dum RP, Strick PL. 1993. Topographic organization of corticospinal projections from the frontal lobe: motor areas on the lateral surface of the hemisphere. J Neurosci 13(3):952-80.
He SQ, Dum RP, Strick PL. 1995. Topographic organization of corticospinal projections from the frontal lobe: motor areas on the medial surface of the hemisphere. J Neurosci 15(5 Pt 1):3284-306.
Heffner RS, Masterton RB. 1983. The role of the corticospinal tract in the evolution of human digital dexterity. Brain Behav Evol 23(3-4):165-83.
Hendrix CM, Mason CR, Ebner TJ. 2009. Signaling of grasp dimension and grasp force in dorsal premotor cortex and primary motor cortex neurons during reach to grasp in the monkey. J Neurophysiol 102(1):132-45. 163
Henneman E, Somjen G, Carpenter DO. 1965. FUNCTIONAL SIGNIFICANCE OF CELL SIZE IN SPINAL MOTONEURONS. J Neurophysiol 28:560-80.
Herbert WJ, Davidson AG, Buford JA. 2010. Measuring the motor output of the pontomedullary reticular formation in the monkey: do stimulus-triggered averaging and stimulus trains produce comparable results in the upper limbs? Exp Brain Res 203(2):271-83.
Herz J, Reitmeir R, Hagen SI, Reinboth BS, Guo Z, Zechariah A, ElAli A, Doeppner TR, Bacigaluppi M, Pluchino S, Kilic U, Kilic E, Hermann DM. 2012. Intracerebroventricularly delivered VEGF promotes contralesional corticorubral plasticity after focal cerebral ischemia via mechanisms involving anti-inflammatory actions. Neurobiol Dis 45(3):1077-85.
Hikosaka O, Kawakami T. 1976. Inhibitory interneurons in the reticular formation and their relation to vestibular nystagmus. Brain Res 117(3):513-8.
Hikosaka O, Kawakami T. 1977. Inhibitory reticular neurons related to the quick phase of vestibular nystagmus--their location and projection. Exp Brain Res 27(3-4):377-86.
Homma Y, Nonaka S, Matsuyama K, Mori S. 1995. Fastigiofugal projection to the brainstem nuclei in the cat: an anterograde PHA-L tracing study. Neurosci Res 23(1):89- 102.
Hongo T, Jankowska E, Lundberg A. 1965. Effects evoked from the rubrospinal tract in cats. Experientia 21(9):525-6.
Hoshi E, Tanji J. 2004. Differential roles of neuronal activity in the supplementary and presupplementary motor areas: from information retrieval to motor planning and execution. J Neurophysiol 92(6):3482-99.
Houdijk H, ter HN, Nooijen C, Rijntjes D, Tolsma M, Lamoth C. 2010. Energy expenditure of stroke patients during postural control tasks. Gait Posture 32(3):321-6.
Huang MH, Brown SH. 2013. Age differences in the control of postural stability during reaching tasks. Gait Posture.
Hummelsheim H, Wiesendanger M, Bianchetti M, Wiesendanger R, Macpherson J. 1986. Further investigations of the efferent linkage of the supplementary motor area (SMA) with the spinal cord in the monkey. Exp Brain Res 65(1):75-82.
Humphrey DR, Rietz RR. 1976. Cells of origin of corticorubral projections from the arm area of primate motor cortex and their synaptic actions in the red nucleus. Brain Res 110(1):162-9.
164
Jacobs JV, Lou JS, Kraakevik JA, Horak FB. 2009. The supplementary motor area contributes to the timing of the anticipatory postural adjustment during step initiation in participants with and without Parkinson's disease. Neuroscience 164(2):877-85.
Jankowska E. 1975. Cortical motor representation in view of recent experiments on cortico-spinal relations. Acta Neurobiol Exp (Wars ) 35(5-6):699-706.
Jankowska E, Edgley SA. 2006. How can corticospinal tract neurons contribute to ipsilateral movements? A question with implications for recovery of motor functions. Neuroscientist 12(1):67-79.
Jankowska E, Hammar I, Slawinska U, Maleszak K, Edgley SA. 2003. Neuronal basis of crossed actions from the reticular formation on feline hindlimb motoneurons. J Neurosci 23(5):1867-78.
Jankowska E, Padel Y, Tanaka R. 1975. The mode of activation of pyramidal tract cells by intracortical stimuli. J Physiol 249(3):617-36.
Jankowska E, Stecina K. 2007. Uncrossed actions of feline corticospinal tract neurones on lumbar interneurones evoked via ipsilaterally descending pathways. J Physiol 580(Pt 1):133-47.
Jefferson SC, Tester NJ, Howland DR. 2011. Chondroitinase ABC promotes recovery of adaptive limb movements and enhances axonal growth caudal to a spinal hemisection. J Neurosci 31(15):5710-20.
Kably B, Drew T. 1998. Corticoreticular pathways in the cat. I. Projection patterns and collaterization. J Neurophysiol 80(1):389-405.
Kaeser M, Wyss AF, Bashir S, Hamadjida A, Liu Y, Bloch J, Brunet JF, Belhaj-Saif A, Rouiller EM. 2010. Effects of unilateral motor cortex lesion on ipsilesional hand's reach and grasp performance in monkeys: relationship with recovery in the contralesional hand. J Neurophysiol 103(3):1630-45.
Keizer K, Kuypers HG. 1984. Distribution of corticospinal neurons with collaterals to lower brain stem reticular formation in cat. Exp Brain Res 54(1):107-20.
Keizer K, Kuypers HG. 1989. Distribution of corticospinal neurons with collaterals to the lower brain stem reticular formation in monkey (Macaca fascicularis). Exp Brain Res 74(2):311-8.
Kermadi I, Liu Y, Tempini A, Calciati E, Rouiller EM. 1998. Neuronal activity in the primate supplementary motor area and the primary motor cortex in relation to spatio- temporal bimanual coordination. Somatosens Mot Res 15(4):287-308.
165
Kimberley TJ, Khandekar G, Skraba LL, Spencer JA, Van Gorp EA, Walker SR. 2006. Neural substrates for motor imagery in severe hemiparesis. Neurorehabil Neural Repair 20(2):268-77.
King JS, Martin GF, Conner JB. 1972. A light and electron microscopic study of corticorubral projections in the opossum, Didelphis marsupialis virginiana. Brain Res 38(2):251-65.
Kneisley LW, Biber MP, LaVail JH. 1978. A study of the origin of brain stem projections to monkey spinal cord using the retrograde transport method. Exp Neurol 60(1):116-39.
Koch M, Schnitzler HU. 1997. The acoustic startle response in rats--circuits mediating evocation, inhibition and potentiation. Behav Brain Res 89(1-2):35-49.
Kraskov A, Dancause N, Quallo MM, Shepherd S, Lemon RN. 2009. Corticospinal neurons in macaque ventral premotor cortex with mirror properties: a potential mechanism for action suppression? Neuron 64(6):922-30.
Kraskov A, Prabhu G, Quallo MM, Lemon RN, Brochier T. 2011. Ventral premotor- motor cortex interactions in the macaque monkey during grasp: response of single neurons to intracortical microstimulation. J Neurosci 31(24):8812-21.
Kuchler M, Fouad K, Weinmann O, Schwab ME, Raineteau O. 2002. Red nucleus projections to distinct motor neuron pools in the rat spinal cord. J Comp Neurol 448(4):349-59.
Kurata K, Tanji J. 1986. Premotor cortex neurons in macaques: activity before distal and proximal forelimb movements. J Neurosci 6(2):403-11.
Kuypers HG. 1960. Central cortical projections to motor and somato-sensory cell groups. An experimental study in the rhesus monkey. Brain 83:161-84.
Kwakkel G, Kollen B, Wagenaar R. 1999. Therapy Impact on Functional Recovery in Stroke Rehabilitation: A critical review of the literature. 85 ed. p 377-91.
Kwan HC, MacKay WA, Murphy JT, Wong YC. 1978. Spatial organization of precentral cortex in awake primates. II. Motor outputs. J Neurophysiol 41(5):1120-31.
Labriola JE, Lee TQ, Debski RE, McMahon PJ. 2005. Stability and instability of the glenohumeral joint: the role of shoulder muscles. J Shoulder Elbow Surg 14(1 Suppl S):32S-8S.
Lacroix S, Havton LA, McKay H, Yang H, Brant A, Roberts J, Tuszynski MH. 2004. Bilateral corticospinal projections arise from each motor cortex in the macaque monkey: a quantitative study. J Comp Neurol 473(2):147-61.
166
Lawrence DG, Kuypers HG. 1968. The functional organization of the motor system in the monkey. I. The effects of bilateral pyramidal lesions. Brain 91(1):1-14.
Lee S, Yao J, Acosta AM, Dewald JA. 2009. The effects of asymmetric tonic neck reflex during reaching movement following stroke: preliminary results. Conf Proc IEEE Eng Med Biol Soc 2009:1581-4.
Lee YS, Lin CY, Jiang HH, Depaul M, Lin VW, Silver J. 2013. Nerve regeneration restores supraspinal control of bladder function after complete spinal cord injury. J Neurosci 33(26):10591-606.
Lemon RN. 2008a. An enduring map of the motor cortex. Exp Physiol 93(7):798-802.
Lemon RN. 2008b. Descending pathways in motor control. Annu Rev Neurosci 31:195- 218.
Lemon RN, Griffiths J. 2005. Comparing the function of the corticospinal system in different species: organizational differences for motor specialization? Muscle Nerve 32(3):261-79.
Lemon RN, Kirkwood PA, Maier MA, Nakajima K, Nathan P. 2004. Direct and indirect pathways for corticospinal control of upper limb motoneurons in the primate. Prog Brain Res 143:263-79.
Lindgren I, Jonsson AC, Norrving B, Lindgren A. 2007. Shoulder pain after stroke: a prospective population-based study. Stroke 38(2):343-8.
Lingenhohl K, Friauf E. 1992. Giant neurons in the caudal pontine reticular formation receive short latency acoustic input: an intracellular recording and HRP-study in the rat. J Comp Neurol 325(4):473-92.
Liu J, Morel A, Wannier T, Rouiller EM. 2002. Origins of callosal projections to the supplementary motor area (SMA): a direct comparison between pre-SMA and SMA- proper in macaque monkeys. J Comp Neurol 443(1):71-85.
Liu Z, Li Y, Zhang L, Xin H, Cui Y, Hanson LR, Frey WH, Chopp M. 2012. Subacute intranasal administration of tissue plasminogen activator increases functional recovery and axonal remodeling after stroke in rats. Neurobiol Dis 45(2):804-9.
Luppino G, Matelli M, Camarda RM, Gallese V, Rizzolatti G. 1991. Multiple representations of body movements in mesial area 6 and the adjacent cingulate cortex: an intracortical microstimulation study in the macaque monkey. J Comp Neurol 311(4):463- 82.
Macpherson JM, Marangoz C, Miles TS, Wiesendanger M. 1982. Microstimulation of the supplementary motor area (SMA) in the awake monkey. Exp Brain Res 45(3):410-6. 167
Maier MA, Armand J, Kirkwood PA, Yang HW, Davis JN, Lemon RN. 2002a. Differences in the corticospinal projection from primary motor cortex and supplementary motor area to macaque upper limb motoneurons: an anatomical and electrophysiological study. Cereb Cortex 12(3):281-96.
Maier MA, Armand J, Kirkwood PA, Yang HW, Davis JN, Lemon RN. 2002b. Differences in the corticospinal projection from primary motor cortex and supplementary motor area to macaque upper limb motoneurons: an anatomical and electrophysiological study. Cereb Cortex 12(3):281-96.
Maier MA, Bennett KM, Hepp-Reymond MC, Lemon RN. 1993. Contribution of the monkey corticomotoneuronal system to the control of force in precision grip. J Neurophysiol 69(3):772-85.
Marsden CD, Merton PA, Morton HB. 1981. Human postural responses. Brain 104(3):513-34.
Martin GF, Dom R. 1971. Reticulospinal fibers of the opossum, Didelphis virginiana. II. Course, caudal extent and distribution. J Comp Neurol 141(4):467-83.
Martin GF, Humbertson AO, Jr., Laxson LC, Panneton WM, Tschismadia I. 1979. Spinal projections from the mesencephalic and pontine reticular formation in the North American Opossum: a study using axonal transport techniques. J Comp Neurol 187(2):373-99.
Massion J. 1992. Movement, posture and equilibrium: interaction and coordination. Prog Neurobiol 38(1):35-56.
Massion J. 1994. Postural control system. Curr Opin Neurobiol 4(6):877-87.
Matsuyama K, Drew T. 1997. Organization of the projections from the pericruciate cortex to the pontomedullary brainstem of the cat: a study using the anterograde tracer Phaseolus vulgaris-leucoagglutinin. J Comp Neurol 389(4):617-41.
Matsuyama K, Kobayashi Y, Takakusaki K, Mori S, Kimura H. 1993. Termination mode and branching patterns of reticuloreticular and reticulospinal fibers of the nucleus reticularis pontis oralis in the cat: an anterograde PHA-L tracing study. Neurosci Res 17(1):9-21.
Matsuyama K, Mori F, Kuze B, Mori S. 1999. Morphology of single pontine reticulospinal axons in the lumbar enlargement of the cat: a study using the anterograde tracer PHA-L. J Comp Neurol 410(3):413-30.
168
Matsuyama K, Takakusaki K, Nakajima K, Mori S. 1997. Multi-segmental innervation of single pontine reticulospinal axons in the cervico-thoracic region of the cat: anterograde PHA-L tracing study. J Comp Neurol 377(2):234-50.
McNeal DW, Darling WG, Ge J, Stilwell-Morecraft KS, Solon KM, Hynes SM, Pizzimenti MA, Rotella DL, Vanadurongvan T, Morecraft RJ. 2010. Selective long-term reorganization of the corticospinal projection from the supplementary motor cortex following recovery from lateral motor cortex injury. J Comp Neurol 518(5):586-621.
Mintzopoulos D, Astrakas LG, Khanicheh A, Konstas AA, Singhal A, Moskowitz MA, Rosen BR, Tzika AA. 2009. Connectivity alterations assessed by combining fMRI and MR-compatible hand robots in chronic stroke. Neuroimage 47 Suppl 2:T90-T97.
Mitz AR, Wise SP. 1987. The somatotopic organization of the supplementary motor area: intracortical microstimulation mapping. J Neurosci 7(4):1010-21.
Mokin M, Keifer J. 2006. Quantitative analysis of immunofluorescent punctate staining of synaptically localized proteins using confocal microscopy and stereology. J Neurosci Methods 157(2):218-24.
Monakow KH, Akert K, Kunzle H. 1979. Projections of precentral and premotor cortex to the red nucleus and other midbrain areas in Macaca fascicularis. Exp Brain Res 34(1):91-105.
Montgomery LR, Herbert WJ, Buford JA. 2013. Recruitment of ipsilateral and contralateral upper limb muscles following stimulation of the cortical motor areas in the monkey. Exp Brain Res.
Montgomery LR, Moran SL, Herbert WJ, Buford JA. 2010. Ipsilateral muscle activity in the upper limb revealed by EMG following motor cortical stimulation using stimulus trains. Society for Neuroscience Program No 492 12.
Morecraft RJ, Herrick JL, Stilwell-Morecraft KS, Louie JL, Schroeder CM, Ottenbacher JG, Schoolfield MW. 2002. Localization of arm representation in the corona radiata and internal capsule in the non-human primate. Brain 125(Pt 1):176-98.
Morecraft RJ, McNeal DW, Stilwell-Morecraft KS, Gedney M, Ge J, Schroeder CM, van Hoesen GW. 2007. Amygdala interconnections with the cingulate motor cortex in the rhesus monkey. J Comp Neurol 500(1):134-65.
Morgan JM, Navabi H, Jasani B. 1997. Role of calcium chelation in high-temperature antigen retrieval at different pH values. J Pathol 182(2):233-7.
169
Morgan JM, Navabi H, Schmid KW, Jasani B. 1994. Possible role of tissue-bound calcium ions in citrate-mediated high-temperature antigen retrieval. J Pathol 174(4):301- 7.
Mori S, Iwakiri H, Homma Y, Yokoyama T, Matsuyama K. 1995. Neuroanatomical and neurophysiological bases of postural control. Adv Neurol 67:289-303.
Mori S, Matsui T, Kuze B, Asanome M, Nakajima K, Matsuyama K. 1998. Cerebellar- induced locomotion: reticulospinal control of spinal rhythm generating mechanism in cats. Ann N Y Acad Sci 860:94-105.
Mushiake H, Inase M, Tanji J. 1991. Neuronal activity in the primate premotor, supplementary, and precentral motor cortex during visually guided and internally determined sequential movements. J Neurophysiol 66(3):705-18.
Nakajima K, Maier MA, Kirkwood PA, Lemon RN. 2000. Striking differences in transmission of corticospinal excitation to upper limb motoneurons in two primate species. J Neurophysiol 84(2):698-709.
Newman DB. 1985a. Distinguishing rat brainstem reticulospinal nuclei by their neuronal morphology. I. Medullary nuclei. J Hirnforsch 26(2):187-226.
Newman DB. 1985b. Distinguishing rat brainstem reticulospinal nuclei by their neuronal morphology. II. Pontine and mesencephalic nuclei. J Hirnforsch 26(4):385-418.
Newman DB, Hilleary SK, Ginsberg CY. 1989. Nuclear terminations of corticoreticular fiber systems in rats. Brain Behav Evol 34(4):223-64.
Otsuka N, Miyashita K, Krieger DW, Naritomi H. 2013. Compensatory contribution of the contralateral pyramidal tract after stroke. Front Neurol Neurosci 32:45-53.
Pandian S, Arya KN. 2012. Relation between the upper extremity synergistic movement components and its implication for motor recovery in poststroke hemiparesis. Top Stroke Rehabil 19(6):545-55.
Patton HD. 1982. Spinal reflexes and synaptic transmission. Excitable Tissues and Reflex Control. Philadelphia, PA: W.B Saunders.
Paxinos G, Huang XF, Toga AW. 2000. The Rhesus Monkey Brain in Stereotaxic Coordinates. San Diego: Academic Press.
Penfield W. 1954. Mechanisms of voluntary movement. Brain 77(1):1-17.
Penfield W, Boldrey E. 1937. Somatic Motor and Sensory Representation in the Cerebral Cortex of Man as Studied by Electrical Stimulation. Brain 60(4):389-443.
170
Pertoldi S, Di BP. 2005. Shoulder-hand syndrome after stroke. A complex regional pain syndrome. Eura Medicophys 41(4):283-92.
Peterson BW. 1979. Reticulospinal projections to spinal motor nuclei. Annu Rev Physiol 41:127-40.
Peterson BW, Anderson ME, Filion M. 1974. Responses of ponto-medullary reticular neurons to cortical, tectal and cutaneous stimuli. Exp Brain Res 21(1):19-44.
Peterson BW, Maunz RA, Pitts NG, Mackel RG. 1975. Patterns of projection and braching of reticulospinal neurons. Exp Brain Res 23(4):333-51.
Peterson BW, Pitts NG, Fukushima K. 1979. Reticulospinal connections with limb and axial motoneurons. Exp Brain Res 36(1):1-20.
Peterson BW, Pitts NG, Fukushima K, Mackel R. 1978. Reticulospinal excitation and inhibition of neck motoneurons. Exp Brain Res 32(4):471-89.
Prentice SD, Drew T. 2001. Contributions of the reticulospinal system to the postural adjustments occurring during voluntary gait modifications. J Neurophysiol 85(2):679-98.
Quallo MM, Kraskov A, Lemon RN. 2012. The activity of primary motor cortex corticospinal neurons during tool use by macaque monkeys. J Neurosci 32(48):17351-64.
Ralston DD, Ralston HJ, III. 1985. The terminations of corticospinal tract axons in the macaque monkey. J Comp Neurol 242(3):325-37.
Rathelot JA, Strick PL. 2006. Muscle representation in the macaque motor cortex: an anatomical perspective. Proc Natl Acad Sci U S A 103(21):8257-62.
Rehme AK, Eickhoff SB, Rottschy C, Fink GR, Grefkes C. 2012. Activation likelihood estimation meta-analysis of motor-related neural activity after stroke. Neuroimage 59(3):2771-82.
Rehme AK, Fink GR, von Cramon DY, Grefkes C. 2011. The role of the contralesional motor cortex for motor recovery in the early days after stroke assessed with longitudinal FMRI. Cereb Cortex 21(4):756-68.
Rho MJ, Cabana T, Drew T. 1997. Organization of the projections from the pericruciate cortex to the pontomedullary reticular formation of the cat: a quantitative retrograde tracing study. J Comp Neurol 388(2):228-49.
Riddle CN, Edgley SA, Baker SN. 2009. Direct and indirect connections with upper limb motoneurons from the primate reticulospinal tract. J Neurosci 29(15):4993-9.
171
Riecker A, Groschel K, Ackermann H, Schnaudigel S, Kassubek J, Kastrup A. 2010. The role of the unaffected hemisphere in motor recovery after stroke. Hum Brain Mapp 31(7):1017-29.
Rizzolatti G, Camarda R, Fogassi L, Gentilucci M, Luppino G, Matelli M. 1988. Functional organization of inferior area 6 in the macaque monkey. II. Area F5 and the control of distal movements. Exp Brain Res 71(3):491-507.
Rizzolatti G, Fadiga L, Gallese V, Fogassi L. 1996. Premotor cortex and the recognition of motor actions. Brain Res Cogn Brain Res 3(2):131-41.
Rosene DL, Roy NJ, Davis BJ. 1986. A cryoprotection method that facilitates cutting frozen sections of whole monkey brains for histological and histochemical processing without freezing artifact. J Histochem Cytochem 34(10):1301-15.
Rosenzweig ES, Brock JH, Culbertson MD, Lu P, Moseanko R, Edgerton VR, Havton LA, Tuszynski MH. 2009. Extensive spinal decussation and bilateral termination of cervical corticospinal projections in rhesus monkeys. J Comp Neurol 513(2):151-63.
Rosenzweig ES, Courtine G, Jindrich DL, Brock JH, Ferguson AR, Strand SC, Nout YS, Roy RR, Miller DM, Beattie MS, Havton LA, Bresnahan JC, Edgerton VR, Tuszynski MH. 2010. Extensive spontaneous plasticity of corticospinal projections after primate spinal cord injury. Nat Neurosci 13(12):1505-10.
Rothwell JC. 2006. The startle reflex, voluntary movement, and the reticulospinal tract. Suppl Clin Neurophysiol 58:223-31.
Sakai ST, Davidson AG, Buford JA. 2009. Reticulospinal neurons in the pontomedullary reticular formation of the monkey (Macaca fascicularis). Neuroscience 163(4):1158-70.
Sato KC, Tanji J. 1989. Digit-muscle responses evoked from multiple intracortical foci in monkey precentral motor cortex. J Neurophysiol 62(4):959-70.
Scheibel ME, Davies TL, Scheibel AB. 1973. Maturation of reticular dendrites: loss of spines and development of bundles. Exp Neurol 38(2):301-10.
Scheibel ME, Scheibel AB. 1961. On circuit patterns of the brain stem reticular core. Ann N Y Acad Sci 89:857-65.
Scheibel ME, Scheibel AB, MOLLICA A, MORUZZI G. 1955. Convergence and interaction of afferent impulses on single units of reticular formation. J Neurophysiol 18(4):309-31.
Schepens B, Drew T. 2003. Strategies for the integration of posture and movement during reaching in the cat. J Neurophysiol 90(5):3066-86.
172
Schepens B, Drew T. 2004. Independent and convergent signals from the pontomedullary reticular formation contribute to the control of posture and movement during reaching in the cat. J Neurophysiol 92(4):2217-38.
Schepens B, Drew T. 2006. Descending signals from the pontomedullary reticular formation are bilateral, asymmetric, and gated during reaching movements in the cat. J Neurophysiol 96(5):2229-52.
Schieber MH. 1990. How might the motor cortex individuate movements? Trends Neurosci 13(11):440-5.
Schieber MH. 2001. Constraints on somatotopic organization in the primary motor cortex. J Neurophysiol 86(5):2125-43.
Schieber MH, Santello M. 2004. Hand function: peripheral and central constraints on performance. J Appl Physiol 96(6):2293-300.
Schnell SA, Staines WA, Wessendorf MW. 1999. Reduction of lipofuscin-like autofluorescence in fluorescently labeled tissue. J Histochem Cytochem 47(6):719-30.
Shibasaki H, Hallett M. 2006. What is the Bereitschaftspotential? Clin Neurophysiol 117(11):2341-56.
Soechting JF, Buneo CA, Herrmann U, Flanders M. 1995. Moving effortlessly in three dimensions: does Donders' law apply to arm movement? J Neurosci 15(9):6271-80.
Soteropoulos DS, Edgley SA, Baker SN. 2011. Lack of evidence for direct corticospinal contributions to control of the ipsilateral forelimb in monkey. J Neurosci 31(31):11208- 19.
Soteropoulos DS, Williams ER, Baker SN. 2012. Cells in the monkey ponto-medullary reticular formation modulate their activity with slow finger movements. J Physiol 590(Pt 16):4011-27.
Spinks RL, Kraskov A, Brochier T, Umilta MA, Lemon RN. 2008. Selectivity for grasp in local field potential and single neuron activity recorded simultaneously from M1 and F5 in the awake macaque monkey. J Neurosci 28(43):10961-71.
Stapley PJ, Drew T. 2009. The pontomedullary reticular formation contributes to the compensatory postural responses observed following removal of the support surface in the standing cat. J Neurophysiol 101(3):1334-50.
Strick PL. 2002. Stimulating research on motor cortex. Nat Neurosci 5(8):714-5.
173
Sung WH, Wang CP, Chou CL, Chen YC, Chang YC, Tsai PY. 2013. Efficacy of coupling inhibitory and facilitatory repetitive transcranial magnetic stimulation to enhance motor recovery in hemiplegic stroke patients. Stroke 44(5):1375-82.
Szabo J, Cowan WM. 1984. A stereotaxic atlas of the brain of the cynomolgus monkey (Macaca fascicularis). J Comp Neurol 222(2):265-300.
Tanji J. 1985. Comparison of neuronal activities in the monkey supplementary and precentral motor areas. Behav Brain Res 18(2):137-42.
Tanji J, Kurata K. 1979. Neuronal activity in the cortical supplementary motor area related with distal and proximal forelimb movements. Neurosci Lett 12(2-3):201-6.
Tanji J, Kurata K. 1981. Contrasting neuronal activity in the ipsilateral and contralateral supplementary motor areas in relation to a movement of monkey's distal hindlimb. Brain Res 222(1):155-8.
Tanji J, Okano K, Sato KC. 1988. Neuronal activity in cortical motor areas related to ipsilateral, contralateral, and bilateral digit movements of the monkey. J Neurophysiol 60(1):325-43.
Tanji J, Shima K. 1994. Role for supplementary motor area cells in planning several movements ahead. Nature 371(6496):413-6.
Tanji J, Shima K, Mushiake H. 1996. Multiple cortical motor areas and temporal sequencing of movements. Brain Res Cogn Brain Res 5(1-2):117-22.
Torvik A, Brodal A. 1957. The origin of reticulospinal fibers in the cat; an experimental study. Anat Rec 128(1):113-37.
Tsai EC, van Bendegem RL, Hwang SW, Tator CH. 2001. A novel method for simultaneous anterograde and retrograde labeling of spinal cord motor tracts in the same animal. J Histochem Cytochem 49(9):1111-22.
Tsai SY, Papadopoulos CM, Schwab ME, Kartje GL. 2011. Delayed anti-nogo-a therapy improves function after chronic stroke in adult rats. Stroke 42(1):186-90.
Umeda T, Takahashi M, Isa K, Isa T. 2010. Formation of descending pathways mediating cortical command to forelimb motoneurons in neonatally hemidecorticated rats. J Neurophysiol 104(3):1707-16.
Valverde F. 1961. Reticular formation of the pons and medulla oblongata. A Golgi study. J Comp Neurol 116:71-99. van den Brand R, Heutschi J, Barraud Q, DiGiovanna J, Bartholdi K, Huerlimann M, Friedli L, Vollenweider I, Moraud EM, Duis S, Dominici N, Micera S, Musienko P, 174
Courtine G. 2012. Restoring voluntary control of locomotion after paralyzing spinal cord injury. Science 336(6085):1182-5.
Van Gisbergen JA, Robinson DA, Gielen S. 1981. A quantitative analysis of generation of saccadic eye movements by burst neurons. J Neurophysiol 45(3):417-42.
Van Kan PL, McCurdy ML. 2002. Discharge of primate magnocellular red nucleus neurons during reaching to grasp in different spatial locations. Exp Brain Res 142(1):151- 7. van Meer MP, Otte WM, van der Marel K, Nijboer CH, Kavelaars A, van der Sprenkel JW, Viergever MA, Dijkhuizen RM. 2012. Extent of bilateral neuronal network reorganization and functional recovery in relation to stroke severity. J Neurosci 32(13):4495-507. van Meer MP, van der Marel K, Wang K, Otte WM, El BS, Roeling TA, Viergever MA, Berkelbach van der Sprenkel JW, Dijkhuizen RM. 2010. Recovery of sensorimotor function after experimental stroke correlates with restoration of resting-state interhemispheric functional connectivity. J Neurosci 30(11):3964-72.
Wannier T, Deliagina TG, Orlovsky GN, Grillner S. 1998. Differential effects of the reticulospinal system on locomotion in lamprey. J Neurophysiol 80(1):103-12.
Weinrich M, Wise SP. 1982. The premotor cortex of the monkey. J Neurosci 2(9):1329- 45.
Welker WI, Benjamin RM, Miles RC, Woolsey CN. 1957. Motor effects of stimulation of cerebral cortex of squirrel monkey (Saimiri sciureus). J Neurophysiol 20(4):347-64.
West MJ. 1993. New stereological methods for counting neurons. Neurobiol Aging 14(4):275-85.
West MJ, Slomianka L, Gundersen HJ. 1991. Unbiased stereological estimation of the total number of neurons in thesubdivisions of the rat hippocampus using the optical fractionator. Anat Rec 231(4):482-97.
Wiesendanger M, Seguin J, Kunzle H. 1974. The Supplementary Motor Area - a control system for posture. In: Stein R, Pearson K, Smith R, Redford J, editors. Control of Posture and Locomotion. New York: Springer US; p 331-46.
Woolsey CN, CHANG HT. 1947. Cortical origin of the pyramidal tract as defined by antidromic volleys from the medullary pyramid. Fed Proc 6(1 Pt 2):230.
Woolsey CN, Settlage PH, Meyers DR, Sencer W, Pinto HT, Travis AM. 1952. Patterns of localization in precentral and "supplementary" motor areas and their relation to the concept of a premotor area. Res Publ Assoc Res Nerv Ment Dis 30:238-64. 175
Yazawa S, Ikeda A, Kunieda T, Mima T, Nagamine T, Ohara S, Terada K, Taki W, Kimura J, Shibasaki H. 1998. Human supplementary motor area is active in preparation for both voluntary muscle relaxation and contraction: subdural recording of Bereitschaftspotential. Neurosci Lett 244(3):145-8.
Yeomans JS, Frankland PW. 1995. The acoustic startle reflex: neurons and connections. Brain Res Brain Res Rev 21(3):301-14.
Yeomans JS, Li L, Scott BW, Frankland PW. 2002. Tactile, acoustic and vestibular systems sum to elicit the startle reflex. Neurosci Biobehav Rev 26(1):1-11.
Yoshino-Saito K, Nishimura Y, Oishi T, Isa T. 2010. Quantitative inter-segmental and inter-laminar comparison of corticospinal projections from the forelimb area of the primary motor cortex of macaque monkeys. Neuroscience 171(4):1164-79.
Zemlan FP, Pfaff DW. 1979. Topographical organization in medullary reticulospinal systems as demonstrated by the horseradish peroxidase technique. Brain Res 174(1):161- 6.
176