Radiological Studies on Hippocampal Development: Morphological
Radiological studies on hippocampal development. Morphological variants and their relationship to epilepsy.
Dragan Bajic ! " #$ %&'( )0$0 1!02!( 343 56( ( 72!( 8$ 9@( A696 6BC9D E 2 3 E " E F22 GH"! E I"PQ 72 #$ ' "!" R'2Q
S1"( Q A696Q T3" R! U"$ $Q I23" V 2 T2 WQ X! E& 03( 3 "2 00 $$!3( W2 E& 3Q D6 Q Q XRS8 B5@9B99D6Y9A9DY9`Q
!3 E $( 2 2"$ !"! E E$3 2 2"$ !"! '2"2 2 $ 2 2 Q !3 2 "( 2 2"$ !"! ' " 2 2"$! 0 ! E$Q )E "$ ( 2 2"$! 2 E$ "! 3 #Q XE 2 " 2 "$ 2 2"$! $ 2 ! E$Q 72 " " "$ 2"$ GXUXPQ 72 $ E 2 ! ' ! 2 Ea!" E XUX 9" " "2 ! # 2 $ E 2 2"$ 3 !3 $! E!Q I3" " GITP $3 E A69 9D6 9" ! 1" ' ! '2! 0'3 "" Q XUX ' E! 9B b c! E " GA6 E9 @ PQ `6b E 2 A69 2 XUX Gd6 E9( d 329( 9Y PQ 72 EE" ' " 3E" Ge6Q6APQ ADb E 2 $ 2 XUXQ 72 Ea!" ' 3E" 232 2 "Q 72 "! ' $ XUXQ ddb E 2 T" D5b E 2 "3" 3c 2 XUXQ XUX " 3 E ! " $ 2 E 2 Q f !! #$ E 9Y6 $! "2 ' cQ 72 3 #$ ' A`9Ad %g Ad "2( AD9A@ %g 5A "2( AB9`Y %g Yd "2Q XUX ' E! D6b( ADb 9db( "Q 72 Ea!" EE" ' 2 "2 e AD %g h AD %g ' " 3E" Ge 6Q669PQ H$ AD %g '( 2 Ea!" E XUX $ 2 2 ! !Q ITX E Y` E! '2! " 23 ' ' ' 3Q 72 ITX ' E$ 3 '0 G%gP 9599@ Y6 %g 9B9`D 72 2"$ !"! ' ( 9 !( `D E! %g 959`AQ 72 E 2$ ' %g `AQ 72 !"! ' " %g A9 2 ( !( ' E$ %g `` ' Q X AYiY` E! Gd9bP( 2 2"$ $ ' $$" A` E!( 2 32 2 EQ
ITX( S $( U"$!( W
! ! !" # # $ %&'()*+)# % p 3 S1" A696
XRS8 B5@9B99D6Y9A9DY9` !C CC!!C99`9BBY G2Cii!Q0 Qiq!r!C CC!!C99`9BBYP TO MY FAMILY
Ne mogu vratiti vrijeme koje nisam proveo sa vama ali ste zauvijek u mojim mislima i mom srcu.
LIST OF PAPERS
This thesis is based on the following studies, which will be referred to in the text by their Roman numerals.
I. Bajic D, Wang C, Kumlien E, Mattsson P, Lundberg S, Eeg-Olofsson O, Raininko R.: Incomplete inversion of the hippocampus - a common developmental anomaly. Eur Radiol. 2008 Jan;18(1):138-42.
II. Bajic D, Kumlien E, Mattsson P, Lundberg S, Wang C, Raininko R.: Incomplete hippocampal inversion - is there a relation to epilepsy? Eur Radiol. 2009 Oct;19(10):2544-50.
III. Bajic D, Ewald U, Raininko R.: Hippocampal development at gestation weeks 23 to 36. An ultrasound study on preterm neonates. Neuroradiology. 2010 Jun;52(6):489-94.
IV. Bajic D, Canto Moreira N, Wikström J, Raininko R.: Asymmetric hippocampal development is common. A fetal magnetic resonance study. Variation and asymmetry of the fetal hippocampus. Manuscript submitted.
TABLE OF CONTENTS
LIST OF PAPERS ...... 5 TABLE OF CONTENTS ...... 7 ABBREVIATIONS ...... 9 INTRODUCTION ...... 11 Milestones in brain research ...... 11 History of imaging with CT and MRI ...... 11 History of limbic system research ...... 12 History of the hippocampus research...... 12 Limbic lobe/ system, short overview of the anatomy ...... 13 Hippocampal embryology ...... 14 Hippocampal anatomy and functions ...... 16 Hippocampal pathology ...... 18 Imaging in hippocampus assessment ...... 19 US ...... 19 CT ...... 19 MRI ...... 19 PET ...... 19 SPECT ...... 20 MRS ...... 20 Diffusion M RI ...... 20 Introduction to present studies ...... 20 AIMS OF THE STUDY ...... 21 MATERIALS AND METHODS ...... 23 Subjects ...... 23 Children and adults (paper I and II) ...... 23 Premature neonates (paper III) ...... 23 Fetuses (paper IV) ...... 23 MR methodology in children and adults (paper I and II) ...... 24 Prestudy –validation of the MR technique ...... 24 Studies of the children and adults (paper I and II) ...... 24 Ultrasound methodology on premature neonates (Paper III) ...... 26 MR methodology in fetuses (paper IV) ...... 27 Statistical analysis for all studies ...... 28 RESULTS ...... 29 ...... 29 Occurrence of the IHI in the non-epileptic and epileptic populations (Paper II) ...... 30 Control g roup ...... 30 Epilepsy patient groups ...... 30 Laterality and distribution of the total and partial forms of the IHI ...... 30 Control g roup ...... 30 Epilepsy patient groups ...... 30 ...... 32 Ultrasound study on preterm neonates (Paper III) ...... 32 MRI study on fetuses (Paper IV) ...... 32 DISCUSSION ...... 35 Studies on the hippocampal development ...... 35 Studies focusing on the morphological variants of the hippocampus and their relation to some epilepsy syndromes ...... 39 CONCLUSIONS ...... 41 ACKNOWLEDGEMENTS ...... 45 REFERENCES ...... 47 ABBREVIATIONS
CT Computerized tomography FDG Fluorodeoxyglucose GA Gestation age GW Gestation week HIMAL Hippocampal malrotation IHI Incomplete hippocampal inversion LTLE Lateral temporal lobe epilepsy MRI Magnetic resonance imaging MRS Magnetic resonance spectroscopy MTLE Mesial temporal lobe epilepsy MTS Mesial temporal sclerosis NMR Nuclear magnetic resonance PET Positron emission tomography SGA Small for gestational age SPECT Single photon emission computed tomography TLE Temporal lobe epilepsy US Ultrasound GRE Gradient echo
INTRODUCTION
Milestones in brain research 5000 year B.C. Trepanations performed for thera- In 1811 the Scottish surgeon Charles Bell was the peutic purpose. " % ! " ! - sory nerves and claimed that they were anatomi- 4000 year B.C. ! cally separated in the spinal roots [67]. ancient Sumeria. In 1817 James Parkinson published “An essay on 3000 year B.C. The word “Brain,” is used for the the Shaking Palsy”. He described a degenerative " # $ disease of the nervous system that even now is Egyptians however believed that the brain was not known as Parkinson’s disease [60]. " ! ! " "" was discarded while the heart was preserved. It is thought that modern research of the brain started with Paul Broca who discovered a speech 2500 year B.C. The Edwin Smith Surgical Pa- center in 1862 [39]. anatomy of the brain was discovered. The papyrus In 1873 John Hughlings Jackson (1835-1911) pre- documented twenty-six cases of brain injury with ! ! * treatment recommendations [8]. [49]. 450 year B.C. The ancient Greek physician Al- In 1906 Santiago Ramon y Cajal and Camille Gol- cmaeon performed anatomical dissection on ani- gi won the Nobel prize for research of structure mals and concluded that the brain was the central and function of nerve cells [27]. organ of intelligence instead of the heart. In 1929 ' + < &- Hippocrates (460-379, B.C.) also stated that the < \ brain was the site of intelligence. in the human brain. He also developed and dem- onstrated the EEG, the equipment for registration Aristotle (384-322, B.C.) distinguished between of these potentials [79]. long-term and short-term memory. Galen (130-200, A.C.) accepted the Hippocratic History of imaging with CT and MRI approach and as physician to the gladiators was In 1882 Nikola Tesla discovered the rotating mag- conscious of the consequences of brain and spinal !> ! " ! < $ injuries. In 1895 @ " Q ! % In 1543 Andreas Vesalius published one of the X-ray image. % & % $ ' ! - scribed how the nerves and the brain work. He In 1937 Isidor I. Rabi (Nobel prize for Physics dissected both humans and animals and noticed in 1944 " ! " that the brains of many animals and all mammals presence by absorbing or emitting radio waves have the same ventricles as humans. He claimed & ! " that the ventricles could not be the site for emo- !$ tion and memory because animals possessed no In 1946 Edward Mills Purcell published his dis- soul [81]. covery of nuclear magnetic resonance in liquids In 1664 " ! and in solids and independently Felix Bloch pro- book “Cerebri Anatome” was written by Thomas posed the Bloch equations which determine the Willis. time evolution of nuclear magnetization. In 1952
11 Radiological studies on hippocampal development they shared the Nobel Prize for “their develop- History of the hippocampus research ment of new ways and methods for nuclear mag- Ca 1564 Giulio Cesare Aranzi coined the name netic precision measurements”. hippocampus from the Greek word for seahorse In 1971 Raymond Damadian discovered that the &44,.> > ')4,.> % " hydrogen signal in cancerous tissue is different monster). from that of healthy tissue, because tumors con- 1886 Camillo Golgi illustrated the unique and tain more water. complicated hippocampal structure [41]. In 1972 \ ! ' ! ! ^ Q " % 1911 Santiago Ramon y Cajal published “Histolo- invented CT. In 1979 they were awarded the No- gie de Systeme Nerveux” with a drawing of the bel Prize for Medicine. hippocampus and an explanation of possible im- In 1973 _ > ! ! ` pulse direction [19]. image (Nobel Prize for Medicine in 2003). 1934 Rafael Lorente de No explained many hip- Between 1974 and 1976 Q - pocampal cell types and their axonal and dendritic ners were installed. patterns. His terminology is in general use today In 1977 " " ! $ [47]. 1956, 1958 Theodor Blackstad, who was called History of limbic system research “the father of the modern hippocampus histolo- In 1878 + ! " " gy”, presented both internal and external neuronal ”le grand lobe limbique” mainly in the context of connections of the hippocampus [12,13]. comparative anatomical studies [55]. a) Olfactory function of the hippocampus In 1937 James Papez described his anatomical 1890 John Hughlings Jackson and Charles Beevor model of emotion, the Papez circuit. He trans- reported a patient with olfactory sensations dur- " ! + ) " ! " ing the seizure [38]. lobe to the functional system -limbic system. His system included hippocampal formation (subicu- 1947 Alf Brodal supported the theory about the lum), fornix, mammillary bodies, mammillotha- hippocampus’s role in olfaction and reported a lamic tract, anterior thalamic nucleus, genu of the number of arguments against this theory [14,15]. internal capsule, cingulate gyrus, cingulum, para- b) Role of the hippocampus in the emotions hippocampal gyrus, entorhinal cortex, perforant In 1937 Papez explained the role of the limbic sys- > ! ! > tem in the emotions through the construction of again the hippocampus [59]. the special system called the “Papez circuit”. In In 1948 Yakovlev added new structures associ- that circuit, the cortical and subcortical structures ated with emotions (orbitofrontal and insular tem- were connected from the hippocampus to the neo- poral lobe cortex, the amygdala and dorsomedial cortex and back to the hippocampus. This concept thalamic nucleus - Yakovlev´s circuit) [83]. " ! " 1952 Paul D. MacLean included additional struc- responsible for emotions, and according Papez, tures in a more dispersed “limbic system,” The the hippocampus’s role was to collect sensory in- term was formally introduced by MacLean. He formation which in later phases provided the basis assumed that the main function of the limbic sys- for the emotions [59]. tem was the formation and expression of emotions In 1982 Brodal rejected Papez`s theory and con- > ! & ! |}}~$ cluded that his theory “is of historical interest only” p. 672 [15].
12 Introduction c) Role of the hippocampal function in atten- Limbic lobe/ system, short overview tion control of the anatomy In 1938 Richard Jung and Alois Kornmuller dis- Limbic lobe, a supernumerary lobe, [49] an arbi- covered a “theta” wave on EEG in the rabbit hip- trary cortical area [40], is situated in the cerebral pocampus [42]. hemisphere partly in the medial temporal lobe and In 1954 Green and Arduini reported that “theta” partly surrounding the corpus callosum. From an waves were connected with enhanced attention anatomical point of view, its structures are part of [30]. the frontal, parietal and temporal lobe. There are cortical structures: the parahippocampal, cingu- 2000 Bush et al, revealed that the cingulate gyrus late and subcallosal gyrus and the hippocampus, and the hippocampus were responsible for general in the limbic lobe. Papez transformed the anatom- attention control [18]. ! " d) Role of the hippocampus in the memory system the limbic system where some subcortical In 1900 ! " + " ! structures were included. R. Jinkins divided the " " ! ! limbic system into cortical structures (cornu am- of the hippocampus found at autopsy [9]. monis, gyrus dentatus, subiculum, entorhinal cor- tex, subcallosal and supracallosal gyrus, medial/ In 1957 William Scoville and Brenda Milner de- lateral longitudinal stria of Lancisii, cingulated scribed the unexpected outcome of surgery with gyrus, parahippocampal gyrus and paratermi- severe anterograde and partial retrograde amnesia nal gyrus) and subcortical structures (amygdala, after destruction of the hippocampus in an epi- habenular nuclei, septal nuclei, hypothalamus, lepsy patient. The patient was unable to form new anterior thalamic nuclei, epithalamus, mesen- episodic memories and could not remember any cephalic tegmentum, mammillary bodies, and events that occurred just prior to his surgery, but fornices) [40]. Other authors divided the limbic retained memories for things that happened years system structures into two arches, outer and in- earlier [70]. ner, separated by the sulcus of the corpus cal- In 1963-1972 Robert Isaacson, Robert Douglas losum anteriorly and hippocampal sulcus poste- and Daniel Kimble suggested “response inhibi- riorly [45,46]. The main structures of the limbic tion hypothesis”, a widely held theory of hyperac- system being interrelated by a complex system of tivity. Interpretation of this theory includes work- $ " ! < ing memory impairment [38]. the fornix across corpus callosum to the hypo- " ! < ! ! In 1972 Endel Tulving distinguished between "" $ "" & memory for episodes and memory for semantic > items [80]. " > ! "" In 1970-1973 O´Keefe and Nadel independently to the mammillary body and after that, via the developed hippocampus cognitive map theory mammillothalamic tract to the anterior thalamic [58]. > ! < " - bres to the anterior cingulated gyrus. From this > & - riorly and to the hippocampus via the entorhinal cortex posteriorly [59]. When Papez demonstrated his circuit, he believed that it was responsible for the emotions. More recent research presented its role in the memory [51] and postulated that Perez’s theory “is of historical interest only” [15].
13 Radiological studies on hippocampal development
Hippocampal embryology The hippocampal development has not been stud- because of the pressure from the gyrus dentatus ied by a great number of researchers. All these and cornu ammonis and moves over into a com- studies were referred to a small number of abort- pression zone. The parahippocampal gyrus and ed fetuses of different gestational ages and micro- subiculum project more medially than the gyrus scopically analysed [2,35,45,46,65]. Some of them dentatus and cornu ammonis [35,45]. performed an MRI examination of the speciments By the GW 18 the hippocampus is infolded into [45,46,65]. In recent times, there are MRI stud- the temporal lobe. The form of the hippocampal ies of the fetal brain analysing the development formation is rounded with the long axis orientated of some of the hippocampal structure [29,66]. The vertically (Fig. 2a). The parahippocampal gyrus development of the hippocampus starts before and subiculum are larger, the temporal horn is GW 10, where are dentate gyrus and cornu am- smaller and the germinal matrix reduced in the monis situated in the posteromedial wall of the size compared with the previous period. As the lateral ventricle [35] (Fig. 1). dentate gyrus enlarges and rotates inward, the At GW 10 a broad shallow hippocampal sulcus two sides of the hippocampal sulcus narrows and appears on the medial aspect of the temporal lobe, " ! $ @ during thickening of the dentate gyrus. inward rotation of the gyrus dentatus, the diffuse At GW 11 the shallow hippocampal sulcus is lo- zone gradually becomes aligned with the hip- calized between the dentate gyrus and the cornu " ! " " ! ! - ammonis. nally transformed into the compression zone. In At GW 12 the hippocampal sulcus is still shal- the deeper portions of the compression zone the low but pointed toward the cornu ammonis. At the cells progressively degenerate and disappear, as tip of the sulcus the diffuse cellular zone appears. the compression increases, to leave a linear zone Later during the development this zone moves that becomes less and less cellular. Even before the deeper into the temporal lobe and the hippocam- diffuse zone is in alignment with the hippocam- pal sulcus follows it. pal sulcus, fusion has begun as the diffuse zone is transformed into the compression zone. [35,45]. At GW 13-14 the dentate gyrus increases in thickness and makes the hippocampal sulcus more Up to GW 21 the hippocampal sulcus and sur- ! ! ! ! > ! ! rounding structures are similar to those found in cornu ammonis. The diffuse cellular zone be- adults (Fig. 2b). The hippocampal sulcus fuses but " " ! ! ! |}~$ sometimes can remain with pia mater and blood < ! ! $ ^ ! - At GW 15-16 the sulcus hippocampalis is best fuse zone narrows progressively it comes com- developed in the temporal region, and inversion of pletely in line both with the fusion zone and the the hippocampal formation into the temporal lobe > ! " has begun. The gyrus dentatus is the largest part zone, deeply. At the same time, the compression of the developing hippocampus and its rotation in- zone increases in extent, because the gyrus denta- ward is the fastest. The cornu ammonis follows tus is larger [2,3,68]. this rotation. The diffuse zone becomes linear
14 Introduction
1 HCS 4
2 3 Week 9 Week 10-11 Week 14 Week 21
Figure 1. Schematic diagram of the fetal development of the hippocampal formation. > Q "" > "> " > 'Q ' Q " $
Figure 2. ' " " \@ ! " " $ } vertical longitudinal axis. " " ! < ! < ! " ! * ! & > weighted IR in the coronal plane).
15 Radiological studies on hippocampal development
Hippocampal anatomy and functions Hippocampal formation includes both gray mat- pocampal body and the collateral sulcus there is a ter (dentate gyrus, cornu ammonis and subiculum) white matter area “collateral white matter”. The ! " < ! " $ ! " ! " ! dentate gyrus consists of three layers. The cornu hippocampal body. The temporal horn is almost ammonis is divided into the 4 segments called ! " " $ CA1 - CA4 (Fig. 3) and contains six layers on mi- Sometimes, the hippocampal form is not ovoid croscopy [25]. but more rounded (Fig. 5a) or pyramidal (Fig. 5b) On coronal MRI images, the normally devel- with its vertical diameter equal or longer than the oped adult hippocampus has an ovoid shape with horizontal. In such cases, the collateral sulcus is the horizontal diameter longer than the vertical vertical, the collateral white mater dislocated lat- (Fig. 4 a-c). The sulcus collateralis is situated un- erally and the lateral part of the temporal horn der the hippocampal formation. Between the hip- " $ < ! < ! !
Figure 3. MR image of the normal adult hippocampus with horizontal longitudinal axis. '\ ' " \ > "> Q^ Q ^"" > \ \ $
16 Introduction
Figure 4. Figure 5. MR image of the normal adult hippocampus Non-ovoid forms of the hippocampus. ^& ! " > ! ! " Q ! " > weighted TSE coronal image. ! " $ " ! " weighted IR coronal image. some patients with severe developmental anoma- Both images perpendicular to the long axis of the hippocampus. lies [4,69] and in epilepsy patients [6,7,10,61] but sometimes in the normal population also [10,16,61]. mechanism of this function is not completely ex- There are a few studies regarding this condition plained yet. The hippocampus has an important in healthy population or patients without develop- role in episodic or autobiographical memory, the mental anomalies and without epilepsy [48]. formation of new memories of experienced events There were three main theories about hippocam- [72] but it is a part of the medial temporal lobe $ > " memory system which is responsible for general role in olfaction, was rejected as false, even though declarative memory. The other memory func- there are still some supporters of this theory. tion of the hippocampus relates to the memory of O’Keefe and Nadel developed the second theory space. There is now, almost universal agreement about the hippocampus`s role in “behavioural in- that spatial coding plays an important role in hip- hibition”. The inhibition theory is currently not pocampal function. If the hippocampus is dam- abandoned but doesn’t have wide support [11]. aged, humans may not remember where they have The third theory, now generally accepted, reveals been and getting lost is one of the most common the role of the hippocampus in the memory but the symptoms of amnesia.
17 Radiological studies on hippocampal development
Hippocampal pathology Aging in healthy people leads to a gradual decline * ! ! " ! of some types of memories including substantial and divided into two groups: loss of neurons in the hippocampus and shrinkage 1: Generalized epileptic seizures considered to of the hippocampus on MRI. Recent studies us- originate at some point within, and rapidly en- ing more precise techniques has challenged those gage bilaterally networks. beliefs, suggesting instead that aging has a limit- 2: Focal seizures considered to originate within ing effect on medial-temporal lobe cytoarchitec- networks limited to one hemisphere, which ture ith [64,75,84]. All elderly people do not reveal may be discreetly localized or more widely dis- hippocampal shrinkage, but those who do tend to tributed. perform less well on some memory tasks [71]. There are a number of different disorders hav- Alzheimer’s disease is strongly connected with ing increased predisposition for seizures [26]. The hippocampal disruption [32] and has a severe im- ! ! * pact on many types of cognition. syndromes are characterized by age at onset or Stress has a negative effect on the hippocam- " ##\ pus which contains high levels of glucocorticoid ! $ ! receptors. Stress-related steroids affect the hip- are grouped as: genetic, structural/metabolic or pocampus by reducing the excitability of some % $ hippocampal neurons, by inhibiting the genesis of the word “cryptogenic” was used instead of “epi- new neurons in the dentate gyrus, and by caus- lepsies of unknown cause” [26]. ing atrophy of dendrites in pyramidal cells of the Temporal lobe epilepsy (TLE) is a form of the CA3 region [53]. These effects are evident in post- focal epilepsy where seizures originate in either traumatic stress disorder, schizophrenia and se- medial or lateral temporal areas. Because of strong vere depression [20] and in Cushing’s syndrome, interconnections the seizure often extends and in- a disorder caused by high levels of cortisol in the volves both areas including neighbouring areas bloodstream [73]. on the same side of the brain as well as contralat- Schizophrenia suffering patients can display eral homologues brain areas. There are two main abnormalities of the cerebral cortex. Hippocam- types of the TLE: mesial temporal lobe epilepsy pal changes such as reduction in the size of the (MTLE), arising in the hippocampus, parahip- hippocampus have also been described, as well as pocampal gyrus and amygdala and, lateral tem- changes in synaptic organization and connectivity poral lobe epilepsy (LTLE), arising in the lateral [33]. It is unclear whether hippocampal alterations neocortex. The abnormality most associated with play any role in causing the psychotic symptoms, mesial TLE is hippocampal sclerosis (scarring) which are the most important feature of schizo- characterized by neuronal cell loss and reactive phrenia, but disturbances in long term memory gliosis in the hippocampal area. The hippocampus are frequently observed in people with schizo- is one of the most electrically excitable parts of phrenia [76]. the brain and one of very few brain regions where Epilepsy is a brain disorder characterized by an new neurons are produced throughout life (neuro- enduring predisposition to generate epileptic sei- genesis) [48]. zures and by the neurobiologic, cognitive, psycho- TLE may be a progressive condition where sei- logical and social consequences of this condition. zures become more resistant to medications over An epileptic seizure is a transient occurrence of time. The association between temporal lobe epi- signs and/or symptoms due to abnormal exces- lepsy and hippocampal sclerosis has been recog- sive or synchronous neuronal activity in the brain nised for over a century, but despite many decades |~$ ! ! of basic and clinical research it is still not possible International League Against Epilepsy (ILAE)
18 Introduction
to assign an arrow of causality. One explanation is US that hippocampal sclerosis may represent a pheno- Cranial US is a useful method in the assessment typically similar manifestation of a heterogeneous of the neonates but there is no data concerning the group of pathologies with diverse pathogenesis, use of US in hippocampal imaging. derived from a complex interplay of numerous factors (including genetic, developmental and en- CT vironmental components) Hopefully prospective CT is not good enough to show anatomical details, neuroimaging studies which can be used to follow but some pathological conditions can be detected: its evolution. Although hippocampal sclerosis is tumors, atrophic or ischemic changes. " "" ! _#> "- mon pathological conditions in TLE, are malfor- MRI mation of cortical development, low grade glio- " mas and vascular/ traumatic lesions. visualization of the hippocampal structure due to LTLE can also be affected by strokes, cortical the great contrast between different tissues. MR malformations, multiple sclerosis plaques, etc., images have to be of high resolution because the and be the cause of seizures from this area. Exact- hippocampus is a fairly small structure. The hip- ly how any of the previously mentioned abnormal- pocampi are cylindrical structures and the MR ities actually cause groups of neurons to generate images has to be acquired perpendicular to the seizure activity is complex and not fully known. hippocampal long axis (i.e. in an oblique coronal Because there can be different lesions, there can plane) to respect spatial orientation [22]. MRI ex- also be different mechanisms of generating a sei- aminations of the histological specimens can be zure. So the temporal lobe epilepsies could be < " ! caused by many etiologic factors. strength and exposition time theoretically can be Rolandic epilepsy is the most common focal epi- unlimited, which is not the case with clinical ex- lepsy syndrome in the pediatric age group. This aminations. “benign childhood epilepsy with centrotemporal spikes” [26] has the excellent prognosis regarding PET the seizures, which usually resolve at puberty, and PET is a nuclear medicine technique which pro- EEG, which will be normalized. duces images of functional brain anatomy using gamma rays emitted indirectly by a positron-emit- Imaging in hippocampus assessment ting radionuclide (tracer). It can measure changes In hippocampus imaging there are 2 groups of in glucose metabolism, and neurotransmitter re- methods: ceptor distribution, which are associated with 1: Morphological, including Ultrasound (US), _#$ ! & \ ! \ " * Computerized tomography (CT) and Magnetic are the most commonly used tracers in epilepsy. resonance imaging (MRI). 2: Functional, including Positron emission tomography (PET), Single photon emissions computed tomography (SPECT), Magnetic resonance Spectroscopy (MRS) and functional MRI (fMRI).
19 Radiological studies on hippocampal development
SPECT SPECT shows blood perfusion in the brain and, in They reported that 12 of the 58 hippocampi had a contrast to other methods, patients could be sub- shape deviating from the normal oval shape. The mitted both to ictal and interictal examinations. criteria of IHI have been poorly evaluated; sym- SPECT is the only method which can reveal in- metry, laterality, form and relationship to other ! ! \ ! * $ #Q deviating features in the temporal lobes are not have a quick uptake in the brain and reach peak known and the occurrence of IHI and its relation effect within 2 minutes after injection. This initial to epilepsy is unclear. tracer uptake remains unchanged after 2 hours. When the change of the general shape and Practically if the radiotracer is injected during the the direction of the longitudinal axis of the hip- seizure, the SPECT examination can be acquired pocampus occur to reach the adult form is not well after 2 hours without changing blood perfusion known. Hevner and Kinney [34] reported that the pattern. cornu ammonis regions had reached a mature, MRS fully rotated position after 18 weeks of gestation. Kier et al. [45,46] found that the infolded rela- MRS is a special technique used for characteri- tionship of the dentate gyrus and cornu ammonis zation of the biochemistry of the tissues. MRS was unchanged in a 24-week-old specimen when is based on the unique radio signals of the non- compared to an 18-week-old specimen. Accord- " " ing to the study of Arnold and Trojanowski [2], and distinct from the frequency of water protons. the hippocampus only increases in volume after It can illustrate decrease of functioning neurons in 20th–21st GW with the increasing of the size of epileptogenic tissue . pyramidal cells in the subiculum and cornu am- Diffusion MRI monis but the hippocampal shape remains un- changed. However, the fusion of the walls of the Diffusion MRI is a technique that can be used to " \@ access the integrity of cerebral tissue. Diffusion ! " |}~$ abnormalities such as increased postictal diffusion The earlier studies have described the microscopi- are an indicator of seizure localization in TLE. cal, not macroscopical development and the mate- Introduction to present studies rials have been very small. Also the shape of the brain can be changed after removal from the skull. The fully inverted hippocampus has an oval con- In different studies, this condition has been " > < - designated as Incomplete hippocampal inversion sion is not always simultaneous or symmetrical. If (IHI) [4], hippocampal malrotation (HIMAL) [6] the inversion is not completed, the hippocampus " " " ! | ~$ retains the round or pyramidal shape. Incomplete We use the term “incomplete hippocampal inver- inversion has been described in patients with epi- sion”. The reason for this choice of terminology lepsy [6,7,10,61] or severe midline malformations is that the hippocampus was, most likely, never [4,69]. Two studies [10,61] also included healthy completely rotated in this condition, therefore it controls with abnormal hippocampal shape in would not be described as “malrotated”. 4–10% of the subjects. Bronen and Cheung [16] studied hippocampal formation in 29 volunteers.
20 Aims of the study
AIMS OF THE STUDY
The goals of the present studies are: To document the frequency, laterality, and symme- To assess hippocampal development by analysing try of IHI in subjects without epilepsy or obvious hippocampal form using cranial US in neonates developmental anomalies of the brain (paper I). born preterm (paper III).
To evaluate if there are relations between the hip- To review the development of the hippocampal re- pocampal form and type of epileptic syndromes gion in the fetuses using MRI (paper IV). by comparing the occurrence of IHI in epileptic populations and non-epileptic populations with neither obvious developmental anomalies of the brain nor any pathological conditions affecting the temporal lobes (paper II).
21 Radiological studies on hippocampal development
22 Materials and methods
MATERIALS AND METHODS
Subjects Children and adults (paper I and II) Premature neonates (paper III) Three hundred patients, subjected in MRI exami- Cranial US images of about 400 premature neo- nation, were drawn from the regional epilepsy nates born at Uppsala university hospital and at register. 78 subjects with conditions that could regional hospitals gravitating to the Uppsala uni- affect the temporal lobes such as obvious in- versity hospital in 2006 – 2008 were reviewed. tracranial developmental anomalies, intracranial Premature neonates with intracranial haemor- masses, hydrocephalus or mesial temporal sclero- rhages grade II-IV, ventriculomegaly, massive sis together with 21 subjects with incomplete data periventricular leukomalacia (PVL) or intracra- were excluded from the study. 150 subjects were nial expansive processes and those cases in which used as controls: 34 were healthy volunteers, 91 the examination was technically faulty were ex- patients selected from the hospital archives and 25 ! !$ " ! patients enrolled in the study. In patients included the cranial US, and infants with small germinal > ! matrix haemorrhage (grade I) or small PVL out- the brain were not expected. These patients were side the temporal lobes were included. suffering from common headaches without neuro- Finally selected population consisted of 160 ne- logical symptoms, from cranial nerve symptoms, onates, 86 of them were boys and 74 girls, with or from predominantly spinal symptoms. They gestational ages (GA) of 22 to 34 weeks (mean 27 had no obvious developmental anomalies and did weeks) at delivery. Mean birth weight was 1021 g not have epilepsy. (range 434 g - 2695 g). Examinations of the epilepsy patients and con- trol subjects were mixed by an administrator and Fetuses (paper IV) then analysed by experienced radiologists blinded The brain images from post mortem MRI of 12 to clinical data. aborted fetuses and from 306 clinical fetal in vivo When 50 men and 50 women among the con- MRI were evaluated. trols were analysed the data was evaluated to re- The fetuses that underwent post mortem exami- veal frequency and type of the IHI in this group. nations had been aborted for extracranial pathol- Subjects were aged 9 months to 73 years, with a ogy. The fetuses in which central nervous system mean age of 29 years (75% were between 10 and (CNS) pathology was revealed in autopsy or MRI 54 years). At this time point, the control subjects were excluded. included 20 healthy volunteers, 17 patients en- The clinical fetal MR examinations were per- rolled in the study, as well as 63 patients that were formed in Uppsala, Sweden and in Porto, Portu- selected from the hospital archives. gal. The indications were suspected body mal- From the 300 epilepsy patients, examined with formations, central nervous system abnormalities > & ! !$ " - suspected on previous ultrasound or abnormal rial consisted of 201 epilepsy patients (100 female, screening laboratory tests. Subjects in which re- 101 male), aged 2 months to 74 years, mean age 26 peated ultrasound or clinical studies revealed years (75% were 6–50 years old); and 150 controls brain pathology were excluded from the study. (75 female, 75 male), aged 2 months to 75 years, ! < ! mean age 31 years (75% were 11–54 years old). & ! !$ " "" ! was a widening of the lateral ventricles.
23 Radiological studies on hippocampal development
Fetuses with the atria wider than 10 mm were $ > " > ! excluded [1,24,31,82]. Technically poor examina- ! ! " tions and all examinations in which the hippocam- protocol examinations, and we decided to exclude pal regions were not well evaluable bilaterally or this variable from further analyses. We concluded the coronal slices were asymmetrically positioned > ! > ! were excluded (180 fetuses, all examined in utero). be used for evaluation of the hippocampal regions. Three post mortem examinations and 60 in Studies of the children and adults (paper I utero & " ! and II) analysis. All the examinations both with the routine proto- MR methodology in children and col and the epilepsy protocol, were analysed by one adults (paper I and II) radiologist according to the variables 1–7 in Table 2. If all the coronal slices through the hippocam- Prestudy –validation of the MR technique pus (the head could not be evaluated) showed a Before we began the study we created criteria for deviating shape, we categorized the IHI as total. If the studies and validated MR techniques. We test- ! " ! ! ed for the type of MR examination needed to eval- others deviating, the IHI was of partial type. The uate the structures of the hippocampal region. The images were analysed without knowledge of clini- images of 30 individuals from the hospital archive cal data. The MR examinations were performed were evaluated. All of them were examined with using two techniques (Table 1). A total of 142 epi- two different MR techniques, “a routine protocol” lepsy patients were examined according to an epi- and “an epilepsy protocol”, using thin slices per- lepsy protocol. Fifty-nine patients were examined pendicular to the long axis of the hippocampus. using a routine protocol as they were successfully Both protocols are presented in Table 1. All the treated without a further evaluation according to epilepsy protocol examinations were made at 1.5 $ T, and the routine protocol examinations at 1.5 T in the control group were examined in accordance 22 patients and at 0.5 T in 8 patients. with the epilepsy protocol and 91 in accordance Two experienced radiologists analysed the im- with the routine protocol. ages independently and separately for the two pro- Clinical records of the patients, including EEG, tocols, using the previously established variables % ! ! used for analysis of the hippocampal region (Table of epilepsy and location of the epileptic focus. 2). # ! ! One of the radiologists repeated the evaluations " ! " after 6 months. One radiologist found incomplete and related disorders by the International League inversion in 16 of the 60 hippocampi, the other Against Epilepsy (ILAE) from 1989. in 17 of 60. Intraobserver agreement for variables For these studies the research plan was accepted 1–7 was 97% for epilepsy protocol, and 98% for by the National Ethics Committee. All the volun- routine protocol. Interobserver agreement was teers and the patients included in the prospective 97% for epilepsy protocol and 95% for routine study gave their informed consent.
24 Materials and methods
Table 1. Technique of coronal MR slices
Epilepsy protocol Routine protocol
1.5 T Field strength 1.5 T 0.5 T (in the minority)
Slice thickness 2.5 mm 5 mm
Gap 0.5 mm 0.5 mm
Coronal slice adjusting Perpendicular to the hippocampus In the plane of the posterior pontine long axis contour
Sequences types T2-w TSE, T1-w IR, FLAIR T2-w TSE T1-w SE (not in all)
Abbreviations: T1-w = T1 weighted T2-w = T2 weighted SE = spin echo TSE = turbo spin echo IR = inversion recovery FLAIR = fluid attenuated inversion recovery
Table 2. Variables used for analysis of hippocampal region
In incomplete hippocampal Variable Normally
1. Hippocampal shape Ovoid Round or pyramidal Lateral part of the choroidal 2. Localization of HC Medial part of the choroidal fissure fissure 3. Craniocaudal diameter of the Shorter than the transversal Equal or longer than the transversal HC diameter diameter
4. Tip of the temporal horn Lateral part filled by the HC Lateral part empty and looks large
5. Orientation of the collateral Horizontal or with an angle <70 More vertical sulcus degrees to HC
6. Position of the collateral Under the HC Lateral to the HC white matter*
7. Long axis of the choroidal Horizontal, parallel to HC long Skew or vertical fissure axis
8. Position of the fimbria Cranial to HC Medial to HC
Abbreviations: HC = hippocampus * Collateral white matter is the white matter between the collateral sulcus grey matter and hippocampus just medial to the collateral eminence [13]
25 Radiological studies on hippocampal development
Ultrasound methodology on premature neonates (Paper III) All examinations were performed in the Neona- Table 3. Gestation ages at delivery and tal intensive care unit with an Acuson Sequoia at ultrasound examination US scanner using a high frequency transducer 10 Number of neonates MHz. US examination was a part of the routine Gestations age neurological assessment of the premature neo- (Weeks) At delivery At examination nates. An anterior fontanel approach with stand- 22 3 0 ard six coronal or paracoronal slices in addition to < !$ 23 13 8 experienced paediatric radiologists performed 24 22 16 examinations and the hippocampal forms on the 25 24 19 images was subsequently analysed by one ex- perienced paediatric radiologist. Coronal slices 26 24 22 of the routine cranial US performed at 1st – 7th 27 9 20 postnatal days, preferably on day 1 were analysed. 28 13 9 Gestational ages upon examination are shown in Table 3. 29 17 20 One of the parents of each child gave written 30 14 8 permission to use the images in the study. 31 89 " ! ! ! ! 32 78 slice III or IV at the level of the third ventricle 33 25 (Fig. 6). The hippocampal form was evaluated and the ratio between the horizontal and vertical di- 34 22 ameters of the hippocampal body was calculated. 35 09 The ratios were divided in three categories: 36 0 3 0 ' " $ Total 158 158 1 $} " $ 3: If the ratio was between 1 and 1.5, additional criteria [10,11] such as the enlargement of the lateral part of the temporal horn and the orien- tation of the collateral sulcus were used to de- cide if hippocampal inversion was completed or not.
26 Materials and methods
Figure 6. Ultrasound image of the both hippocampal regions. An ovoid hippocampus on the right side (white arrow) and non-ovoid on the left side (empty arrow).
MR methodology in fetuses (paper IV) All examinations were performed with Philips in the sagittal, axial and coronal planes. The slice Intera MR imagers (Best, The Netherlands) op- thickness was 3-4 mm, and the matrix between erating at 1.5 T. The post mortem examinations 256x256 and 364x256. Fields of view ranged from were performed using a birdcage knee-foot coil. 20x20 to 30x30 cm. No sedation was given to the T2-weighted 2D fast spin echo (FSE) sequences mothers. were performed in the sagittal, coronal and trans- Hippocampal regions were evaluated in one or verse planes, with slice thickness of 2 mm and more coronal slices through the area. Progress in-plane resolution of 0.50x0.53 mm (transverse of the hippocampal inversion was analysed and plane) and 0.59x0.62 mm (coronal and sagittal a comparison between the right and left side was planes). A T1-weighted magnetization-prepared made in every case. In cases where the hippocam- 3D gradient echo (GRE) sequence was performed pal sulcus was not yet closed, the angle between in the sagittal plane, with slice thickness of 1 mm the upper and lower lip of the hippocampal sul- and in-plane resolution 0.59x0.91 mm. From the cus was evaluated. If the hippocampal sulcus was acquired T1-weighted images, reformats were closed the shape of the infolded hippocampus was constructed in the axial and coronal planes. The ! < ! !> " ! in utero studies were performed using abdominal ovoid with a vertical long axis) or ovoid with a ! \ & $ " horizontal long axis. The existence and orienta- included T2-weighted single-shot FSE sequences tion of the collateral sulcus were also recorded.
27 Radiological studies on hippocampal development
One experienced radiologist evaluated all the Statistical analysis for all studies 318 examinations and made exclusions according The statistical packages JMP (version 5.1) and to the criteria presented in the material descrip- StatsDirect 2.5.7 were used. Differences in pro- $ & " ! ! % ! 0 analysis (63 examinations, 126 hippocampal re- method and by Fisher’s (generalised) exact test gions) were then also evaluated by another expe- when expected numbers were small. P values less rienced radiologist independently. Both radiolo- $ } ! ! ¡ }¢ - gists made analyses blinded for gestational ages. ! < !$ ^ Inter-observer agreement in evaluation of the hip- tailed. pocampal structure was 90% and in evaluation of the collateral sulcus 97%. In cases when there was disagreement, the radiologists made reevaluations and consensus was reached. The study protocol was accepted by the local Ethical Committees in Sweden and in Portugal. In Sweden, all mothers gave written informed consent before the examinations. In Portugal, the Ethical Committee approved that the clinical studies could be used in the project without asking mothers for consent.
28 Results
RESULTS
Morphological MR analysis in the (Paper I) A round or pyramidal hippocampal shape was ! ! ! ¢ }¢ ! - left. A vertical collateral sulcus was also found in terval (95% CI) 11.3–26.7%). In all these cases, 17/81 subjects (in 10/40 men and in 7/41 women) the variables 2–7 (Table 1) were also deviating. with normal oval hippocampal shape. The orien- There were no differences between the volunteers tation could differ from the anterior to the poste- and the patients. The frequency was the same in rior part. Defects in the corpus callosum or other men and women (Table 4). developmental brain anomalies were not revealed All the unilateral malformations were found in any of the subjects. on the left side. Total IHI was reported in 9 of 13 Two aspects of the IHI were analysed in paper left-sided IHI and the remaining 4 were partial. two: occurrence and laterality and distribution of Among 6 bilateral IHI four were total bilaterally, the total and partial IHI forms. one total on the left side, partial on the right and
Table 4. Number of subjects with incomplete hippocampal inversion
Men Women Total N=50 N=50 N=100
Subjects with incomplete 10 9 19 hippocampal inversion
Unilateral incomplete inversion 9 4 13 (all on the left side) Total 7 2 9 Partial 2 2 4
Bilateral incomplete 156 hippocampal inversion Total bilaterally 1 3 4 Total on the left 1 1 Total on the right 1 1
29 Radiological studies on hippocampal development
Occurrence of the IHI in the non- Laterality and distribution of the epileptic and epileptic populations total and partial forms of the IHI (Paper II) Control group Control group Of the persons with IHI, 71% had an unilateral IHI including a round or pyramidal hippocampal IHI and 29% had a bilateral IHI. All the unilat- shape combined with presence of all other devi- eral IHIs (20 subjects) were found on the left side: ating variables previously named was found in 13/20 unilateral IHIs were of total type (65%) 28/150 (19%) subjects (Table 5). and 7/20 were partial (35%). Bilateral IHI, found in 8/150 of all control subjects, was total in 5/8 Epilepsy patient groups (63%) subjects and partial at least on one side in ' - 3/8 (37%) (Table 6). ly larger in patients with epilepsy than in controls Epilepsy patient groups (P<0.02) (Table 5). The proportion of the patients ' All epilepsy patients $ > ! " - The distribution of the laterality of IHI did not dif- _# $ > " ! fer between controls and epilepsy patients (Fish- the controls. There were no differences in the age R ! & > $} " - of onset or in therapy refractoriness between the lepsy patient subgroups (codes 1.1, 1.2, 1.3, 2.1, 2.2 groups having or not having IHI. In this group, ! $> R ! & > $ ¥ $ Rolandic epilepsy differed, with a higher propor- In subgroups 2.2 (generalised epilepsy syndromes ' $ $ - cryptogenic) and 3 (undetermined), there was a ' > with generalised epilepsy in comparison to con- bilateral IHI compared with both controls and oth- trols (P<0.01). Notably, the distribution of IHI er epilepsy patients. The distribution of total and among the subtypes of generalised epilepsies (2.1, partial IHI did not differ in controls and epilepsy 2.2 and 2.4) was not homogeneous (Fisher’s gen- patients with the exception of cryptogenic gener- ! & > $ $ ! alised epilepsy, in which all bilateral IHIs (7/21) high proportion of IHI among patients with cryp- were total. togenic generalised epilepsies (code 2.2) and spe- Generalised cryptogenic epilepsy ! " ! $ $ ` ! - pathic generalised epilepsy had an IHI. There was Twenty-one patients had generalised cryptogenic ! ' epilepsy syndrome, 12 of them had IHI. Of these, in patients with epilepsy of undetermined status 57% (7/12) were bilateral. The frequency of bilat- " ! > 0$ > £ $ ' $ ¥ " ! ! R & > $ }} $ ! ! $ ^ - tients in this subgroup who had bilateral IHI had the total form (7/7). The frequency of total IHI was highest in this subgroup compared to both controls and other epilepsy subgroups. Cognitive defects were common both in patients with IHI (12/12) and without IHI (7/9). Therapy refractori- ness was similar in patients with IHI (9/12) and without IHI (6/9) (Table 6).
30 Results
Table 5. Occurrence of IHI in epilepsy syndromes
Subjects with Group No. of subjects P value * IHI N (%)
Controls 150 28 (19) Epilepsy patients (all) 201 60 (30) 0.017 1 Focal epilepsy syndromes 163 44 (27) 0.08 1.1 Idiopathic 31 13 (42) 0.004 Rolandic 25 11 (44) 0.004 Other 6 2 (33) 0.33 1.2 Symptomatic 127 27 (21) 0.59 Temporal 57 14 (25) 0.34 Frontal 7 2 (29) 0.62 Other* 63 11 (18) 0.83 1.3 Cryptogenic 5 4 (80) 0.0062 2 Generalised epilepsy syndrome 33 13 (39) 0.009 2.1 Idiopathic 10 0 0.212 2.2 Cryptogenic 21 12 (57) 0.0001 2.4 Specific syndromes 2 1 (50) 0.346 3 Undetermined (all types) 5 3 (60) 0.0549
* = parietal lobe epilepsies, occipital lobe epilepsies, chronic progressive epilepsia partialis continua of childhood, syndromes characterized by seizures with specific modes of precipitation [23]
Table 6. Distribution of the laterality of IHI in epilepsy syndromes
Subjects with Subjects with Subjects with Subjects with Group IHI N unilateral IHI bilateral IHI unilateral IHI – left N (%) N (%) – right N (%) Controls 28 20 (71) 8 (29) 0 Epilepsy patients (all) 60 40 (67) 16 (27) 4 (7) 1 Focal epilepsy syndromes 44 33 (75) 7 (16) 4 (9) 1.1 Idiopathic 13 10 (77) 2 (15) 1 (8) Rolandic 11 9 (82) 1 (9) 1 (9) Other 2 1 (50) 1 (50) 0 1.2 Symptomatic 27 20 (74) 4 (15) 3 (11) Temporal 14 11 (79) 1 (7) 2 (14) Frontal 2 1 (50) 1 (50) 0 Other** 11 8 (78) 2 (18) 1 (9) 1.3 Cryptogenic 4 3 (75) 1 (25) 0 2 Generalised epilepsy syndrome 13 6 (46) 7 (54) 0 2.1 Idiopathic 0000 2.2 Cryptogenic 12 5 (42) 7 (58) 0 2.4 Specific syndromes 1 1 (100) 0 0 3 Undetermined (all types) 3 1 (33) 2 (67) 0
* = parietal lobe epilepsies, occipital lobe epilepsies, chronic progressive epilepsia partialis continua of childhood, syndromes characterized by seizures with specific modes of precipitation [23]
31 Radiological studies on hippocampal development
Laterality of IHI in relation to EEG Laterality of the IHI was compared with focus ed right sided IHI and isolated left sided IHI was laterality on EEG. Among 14 TLE patients with $ > 0 $ $ IHI, 11 had isolated left-sided, 2 had isolated right-sided and 1 had bilateral IHI. On EEG, the MRI study on fetuses (Paper IV) ' ! ! ! > The three post mortem examinations were per- and both patients with IHI on the right side had formed at GW 17-18 and the 60 in vivo examina- an EEG focus on the left side. Among the remain- \@ $ ^ ! ! ing 11 patients with IHI on the left side, 8 had an hippocampal sulcus bilaterally. The hippocampal EEG focus on the left side, 2 had a bilateral focus sulcus was completely or widely open, bilater- and data for 1 patient were missing. Among 43 ally or unilaterally, in 39/51 fetuses examined at TLE patients without IHI, 19 had an EEG focus GW 17-32 (Fig. 8). In 21/35 fetuses with a bilat- on the left side, 9 on the right side, 2 bilaterally, eral open hippocampal sulcus, the angle between and for the remaining 13, data were not complete. the upper and lower lip, was less than 90 degrees Occurrence of left-sided EEG focus was higher ! $ < > than right-sided in both groups: in 62% of patients than 90 degrees on the right side but more than having IHI and in 63% of patients without IHI. 90 degrees on the left side. The hippocampal sul- cus was closed at GW 21 at the earliest (right side) Ultrasound study on preterm and always bilaterally closed from GW 33 on- neonates (Paper III) wards. Asymmetric development was observed in " 0 1 $} " $ - 11 fetuses in which both hippocampal sulci were tribution of the hippocampal forms (complete open but the angles between the upper and lower versus incomplete inversed) at different ages is lips were different on the right and left sides. The ¥$ < ! more closed hippocampal sulcus was on the right in IHI frequencies between GW 24 and 25. The side in 8 fetuses and on the left side in 3 fetuses. patients are divided into three age groups accord- A closed hippocampal sulcus was found in 28 ing to the IHI frequency. The frequency of the IHI fetuses (Fig. 8). One fetus at GW 21 had a closed was the highest in the group I of infants examined sulcus on the left side only. Three fetuses, at GW at GWs 23-24 12 out of 24 (50%). In the group II 27-30, had a closed right-sided sulcus but the left- GWs 25-28 frequency was 18 out of 72 (25%). In sided sulcus was still open. From GW 33 onwards, the group III GWs 29-36 frequency was 9 out of all hippocampal sulci were closed. ¢ ! The youngest fetus with an open hippocampal frequency of the IHI between group I and group II sulcus on one side and a closed sulcus on the oth- $ ! ! $ > er side showed a non-ovoid shape of the inverted ! $¥} $ hippocampus but the three older fetuses, showed Also frequency of the bilateral IHI was higher an ovoid shape with a horizontal long axis. The in the youngest group: it was found in 7 of 24 non-ovoid hippocampal shapes were found bilat- children (29%) and in 11 of the other 136 children erally at GW 24-29 or unilaterally with the ovoid ¢ $ ! $ > 0 shape on the contralateral side at GW 31-36. From 9.08). Frequency of isolated left-sided IHI was 4 of GW 33 onwards when all hippocampal sulci were 24 (17%) in 23-24 GW group, and 14 of 136 (10%) closed, the hippocampal shape was non-ovoid on in the other premature newborns. This difference the left side and ovoid on the right side in 4/12 $ ! ! ! ' fetuses (33%). rare: one in every group. In the total material, the Asymmetric hippocampal development was difference between the frequencies of the isolat- found in 26 (41%) fetuses when the non-ovoid hip-
32 Results
" ! ! < - deep collateral sulcus bilaterally. Fourteen of the oped than the ovoid shape. The development was 24 fetuses with a bilateral deep collateral sulcus faster on the right in 22 (85%) fetuses and on the had an uni- or bilateral vertical orientation (58%). left only in four (15%). The collateral sulcus was vertical in 5/26 fetuses Development of the collateral sulcus is shown (19%) on the right side and in 13/25 fetuses (52%) in Figure 9. A shallow collateral sulcus was de- on the left side. This difference was statistically tected at the earliest in one fetus at GW 17 but $ $ another fetus showed no collateral sulci at GW 29. Nine fetuses with a non-ovoid hippocampal The deep collateral sulcus was seen at the earliest shape had a deep collateral sulcus. In six of them, at GW 26 on the right and at GW 27 on the left. the nearby collateral sulcus was vertical. At GW 31 or later, all 17 fetuses had developed a
GW 23 3 5
GW 24 4 4 1 7
GW 25 2 2 17
GW 26 2 3 1 16
GW 27 1 4 15
GW 28 2 1 6
GW 29 2 18
GW 30 2 6
GW 31 1 1 7
GW 32 1 1 6
GW 33 5
GW 34-36 1 13
0 5 10 15 20 25 Bilat. IHI Left IHI Right IHI No IHI Figure 7. Occurrence of IHI in premature newborn according to gestation age at the examination
33 Radiological studies on hippocampal development
Number of fetuses
R L 8 R L 7 R L 6 R L R L R L 5 R L 4 R L R L R L 3 R L R L R L R L 2 R L R L R L R L R L R L 1
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Gestation week
Figure 8. Distribution of the hippocampal shape at different gestational ages. Each square pair describes a fetus. ! > _ ! $ B " > C < ! < ! " > C < ! < ! " > & * " < $
Number of fetuses
R L 8 R L 7 R L 6 R L R L R L 5 R L 4 R L R L R L 3 R L R L R L R L 2 R L R L R L R L R L R L 1
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Gestation week
Figure 9. Collateral sulcus at different gestational ages. Each square pair describes a fetus. ! > _ ! $ Collateral sulcus: B < > ΄ > C < > C * $
34 Discussion
DISCUSSION
Studies on the hippocampal development Our fetal MRI study was the second prenatal MR In all fetuses the hippocampal sulcus was iden- ! ! " ! & ! ! in which sulcal development in the hippocampal the previously published data [28,45,46,65] (Ta- region and the shape of the inverted hippocampus ble 7). According to the studies on the formalin- has been evaluated and our ultrasound study of & ! > " ! < ! " ! < - adult appearance at GW 21 [2,3,35,45,46]. We ating hippocampal form using ultrasound. could found a closed hippocampal sulcus only in In the fetal MRI study, we found that there are one fetus at that age. Time variation for the clo- wide individual temporal variations in the deepen- sure is very wide and the majority of the foetus- ing and closure of the hippocampal sulcus. After es had open hippocampal sulci. We found open the closure of the hippocampal sulcus, the invert- hippocampal sulci up to GW 32. In cases with ed hippocampus could be of a non-ovoid shape ! " > ! (Fig. 10) or of an ovoid shape with a near hori- shape of the inverted hippocampal formation us- zontal long axis (Fig. 11). The last-named shape ing the same criteria as in earlier studies on adults, was presumed to present complete hippocampal children and premature neonates. The majority of inversion. the population have ovoid hippocampus with a
Figure 10. Figure 11. In utero MR study. T2-weighted single-shot FSE In utero MR study. T2-weighted single-shot FSE coronal images of fetus at 29 GW. A non-ovoid coronal images of fetus at 29 GW. An ovoid hip- hippocampus with a vertical long axis bilater- pocampus with a closed hippocampal sulcus and a ally with closed hippocampal sulcus. The arrows horizontal long axis bilaterally. The white arrows indicate the hippocampal region and the stars the indicate the hippocampal region, the black closed collateral sulcus. hippocampal sulcus and the star the collateral sulcus.
35 Radiological studies on hippocampal development
Table 7. Detectability of some anatomical structures in the hippocampal region in different ages
Author Bajic et al. Bajic et al. Humphrey Kier et al. Chi et al. Garel et al.
US of the Photos of Fetal MRI Study Autopsy MRI of the Fetal MRI premature the brain (gyration method material specimens neonates (autopsy) study) Number of 63 158 10 6 507 173 subjects Age in 17 - 35 23 - 36 10 - 30 13 - 24 10 - 44 22 - 38 GW Collateral 17 24 23 26 sulcus Open HC 17 - 32 10 13 - 14 22 - 23 sulcus Non ovoid 21 23 15.5 - 18.5 18 - 20 HC shape Ovoid 27 23 20.5 21 HC shape horizontal long axis and that shape we categorize ancy may have been caused by the small numbers as ovoid. All the other shapes were categorized as ! ! non-ovoid. A non-ovoid hippocampus was found " ! $ ! at GW 21 at the earliest and an ovoid hippocampus study only the hippocampal shape was evaluated, at GW 27. Bilateral non-ovoid hippocampi were the hippocampal sulci were not evaluated. In the not observed after GW 29 and bilateral ovoid not fetal MRI study the hippocampal shape was only before GW 29. evaluated if the hippocampal sulcus was closed. In our ultrasound study of preterm neonates, The number of subjects was higher in the US ! ! study. In this study IHI frequency was 14-20% higher frequency of non-ovoid, incomplete in- compared with 19% in a MR study of children versed, hippocampi before GW 25 and the lower and adults. These frequencies are best compara- frequency from GW 25 onwards. Of the neonates ble to the frequency of the uni- or bilateral non- examined up to GW 24, about one half had In- ovoid hippocampi in the fetuses in which both complete hippocampal inversion (IHI). The IHI hippocampal sulci were closed (from GW 33 on- frequency in preterm neonates both at GWs 25– wards). That frequency was 33%. The differences 28 (24%) and at GWs 29–36 (14%) approached the between the frequencies are not statistically sig- IHI frequency (19%) in our MRI studies on chil- $ ^ ! - dren and adults. tal MRI study of the material of 63 fetuses and in The fetal MRI study did not reveal a threshold US study, in the material of 158 infants) the hip- of the frequency of the non-ovoid shape at GW 25 pocampal inversion seems to continue longer than as the ultrasound study had done. This discrep- presumed on the basis of the earlier studies (only
36 Discussion one hippocampus had ovoid shape before GW 29 which indicates the faster hippocampal develop- and about one half had IHI. ment on the right side. Asymmetrical hippocam- In our fetal MRI study even the development pal development was frequent and evenly distrib- of the collateral sulcus was analysed. The collat- uted in different age groups, which were from 17 eral sulcus showed different types of orientation. GW to 36 GW. In a prenatal MR study, where It can be vertical or horizontal which is the more hippocampal infolding angles were measured common type in adults. In 58% of the fetuses with at GW 20-37, no asymmetry was observed [66]. a deep bilateral sulcus, the orientation was ver- However, another group that used the same meth- tical, uni- or bilaterally. The vertical orientation odology in infants and children found a left-right " "" ! $ asymmetry and concluded that it may result from Even the collateral sulcus developed earlier on the slower development of the left hippocampus [57]. ! $ ! ! Chi and co-workers evaluated photographs of 507 early as at GW 17 but was not visible in one fe- " & ! % tus at GW 29. The orientation of the sulcus, hori- age [21]. 207 brains were also sectioned. They de- zontal or vertical, was not related to the gestation scribed the development of many sulci separately age. However, the vertical orientation seems to be but did not mention the hippocampal sulcus in the slightly more common in fetuses than in children report. They found that the right cerebral hemi- and adults. In children and adults, 36% had a ver- sphere, in general, shows gyral complexity earlier tically oriented collateral sulcus uni- or bilaterally. than the left. Even neurophysiological develop- In fetuses with a deep bilateral sulcus, 58% showed ment is slower in the left temporal lobe [62]. that orientation. The difference reached statistical It is not easy to depict small hippocampal struc- $ } $ ! - tures neither on the fetal MRI nor on the US. Im- dren and adults, a non-ovoid hippocampal shape age quality and resolution were the best in the post was always associated with a vertically oriented mortem fetal MRI images (Fig. 12) in which there collateral sulcus. In 3/9 fetuses with a non-ovoid are no movement artefacts and the examination hippocampus and deep collateral sulcus, the near- time is not restricted by the risk of movement. The by sulcus did not have a vertical orientation. This hippocampal region was well assessable in all 12 might be a sign that the hippocampus had not yet post mortem examinations performed at GW 17- reached the ovoid shape but the inversion process 22, included at the start of the study, but only three was continuing. could be accepted in the study owing to brain ab- In our fetal MRI study the number of fetuses normalities in the other cases. In the examinations was higher than in the earlier hippocampal studies in utero, the image quality was much lower (Fig. " & ! ! < 13) and 59% of those examinations were rejected show the wide time variation in the development for technical reasons. The youngest living fetus of the hippocampal region. The timetable of the with a resolution good enough was examined at development of some anatomical structures in the GW 19. Many examinations were rejected as small hippocampal region presented in different studies, details could not be assessable, which is a weak- including ours was shown in Table 7. ness of the study. A general weakness of both our In our fetal MRI study, the development of the fetal MRI study and fetal MRI studies of the for- hippocampus, including the hippocampal sulcus " & ! " " " and the inverted hippocampal formation, was the normal foetuses. An additional weakness of "" ¢ ! ! " & ! " > majority of the asymmetric cases, the right side were possible changes of the anatomical shape }¢ ! ! < ! $ ^ and proportions as a result of the removal of the US study in cases of asymmetric shape of the hip- brains from the skull, placement in the vessels for pocampus the left sided IHI was more common " & !
37 Radiological studies on hippocampal development
Figure 12. Figure 13. Post mortem MR study of a fetus aborted at 17 In utero MR study. T2-weighted single-shot FSE GW. The arrows indicates widely open symmetri- coronal images of fetus at 25 GW. The arrows cal hippocampal sulci. T2-weighted TSE coronal indicate the hippocampal sulci. The hippocampal image. sulcus is larger on the left side. for cutting or imaging. In our ultrasound study as Hippocampal development during the late ges- well as the fetal MRI study the hippocampus was tational period was analysed using cranial ultra- examined in a natural position and shape. sound on preterm neonates. The hippocampal Imaging techniques progress continuously and development in preterm neonates may be af- examination times will be shorter. Images with a fected by pre- and postnatal stress. This problem very high resolution can be obtained even now at exists in all studies on fetal hippocampal devel- ! |¥¥~ opment. Developmental milestones have been limiting factor, in particular, in fetuses which are based on histological and MR studies on forma- more vulnerable than adults. Repeated examina- & ! " % " ! tions would be the best way to demonstrate devel- [2,3,35,45,46,65]. The etiologies of abortions were opmental progress but fetal examinations without not reported. There are no developmental criteria clinical indications are not ethically acceptable. for normal hippocampi in utero because even pre- natal MR studies were performed for medical in- The hippocampal shape was well seen in neo- dications. We have documented the hippocampal nates also with US, especially with high frequency shape in preterm newborns and tried to minimize transducers (10 MHz) but the occurrence of IHI at all those factors possibly affecting hippocampal GWs 29-36 was a little lower than in children and development. Children and adolescents born pre- adults in our MR studies. The difference was not term have been reported to have smaller brain vol- " \ " umes than those born full-term. The volume re- of the partial IHIs might be missed in US exami- duction has been proportionally great particularly nations because brain morphology is not depicted in the hippocampi [36,56,63]. In contrast, in more as well with US as with MRI.
38 Discussion recent samples without major perinatal complica- in healthy volunteers. In our study, 25/200 hip- tions, the hippocampi have shown a marginal, if pocampus (12.5%) were incompletely inverted. any, volume reduction [44,50,52,78]. Intrauterine The frequency of IHI was higher in our subjects growth restriction affects the hippocampal vol- without epilepsy than many other authors have ume [50]. A generally small brain volume and reported in epilepsy patients without corpus cal- cognitive dysfunction consistent with hippocam- losum anomalies. The great variations in fre- pal injury has been found in adolescents born at " ! ! ! ! full-term but SGA [52]. Our study excluded neo- incomplete hippocampal inversion, i.e. including nates SGA. However, there might have been some or excluding of partial incomplete inversions, and unknown factors affecting prenatal hippocampal also on genetic and environmental factors. All development because there is always a cause for incomplete inversions of the hippocampus in our a preterm birth. The time intervals between the study were bilateral or situated on the left side. In birth and the US examination were in our study, our fetal MRI and US studies, the majority of IHI most likely, too short to allow demonstrable post- was also bilateral or on the left side. In the litera- natal developmental deviations. ture, the laterality has not always been mentioned but in the Gamss study, all of the seven deviating Studies focusing on the hippocampal forms were on the left side [28]. morphological variants of the We found that IHI is more common in epilep- hippocampus and their relation to sy patients than in the control subjects group (19 some epilepsy syndromes resp. 30%). The frequency of IHI in our epilepsy In our MRI study, including 100 subjects without patients was higher than in many other studies on obvious developmental anomalies and without epilepsy patients without major brain anomalies. > ! & ! - Barsi et al. [6] described IHI in 6% of 527 patients tion of the IHI as a developmental hippocampal with suspected epilepsy, Peltier et al. [23] in 14% anomaly where the infolding of the hippocampus of 97 epilepsy patients. Baulac et al. [7] found ab- in the temporal lobe was incomplete and the hip- normal hippocampal form in 1.7% (19/1,100) of pocampal form in the MRI coronal slices found patients with partial epilepsy. Thirteen of those 19 to be round or pyramidal. We found that there is patients had isolated hippocampal abnormalities a frequent (19%, 95% CI 11.3–26.7%) hippocam- (2 bilateral, 11 unilateral). In TLE, the frequency pal form variation in the common population, and of IHI ranged from 2.2% (5/222 patients) [10,61] to that this form variation could be total, if the hip- 43% (13/30 patients) but in the last-named study, pocampal shape irregularity is present in all slices 15/30 patients had previously known unilateral on the coronal MR images of the hippocampal re- hippocampal atrophy. In our study, patients with gion; or partial if the shape irregularities are pre- MTS were excluded because of shrunken hip- sent only on some slices. IHI can be bilateral or pocampi are and the original form may not there- unilateral most often on the left side. fore be assessable. There are a few studies regarding IHI in sei- The frequency of IHI was not statistically signif- zure free subjects. One of them is Gamss study icantly higher in focal epilepsies than in controls which was published after our papers I and II [28]. ¥ < $ ¢> $ $ ` > They found rounded hippocampal form and verti- only 25% in TLE. Rolandic epilepsy differs from cal collateral sulcus in 7/497 (1.4%) patients. The the other types of focal epilepsies by having a frequency was lower than in the nonepileptic con- ' trols in the epilepsy studies of Peltier (2/50, 4%) ¢> $ $ ! < " - [61] or Bernasconi (5/50, 10%) [10]. Bronen and tal deviation of the hippocampus in Rolandic epi- Cheung [16,17] reported that 12/58 hippocampi lepsy, although common, may be of milder type (20.8%) had a shape deviating from the normal than in the other epilepsy syndromes. The group
39 Radiological studies on hippocampal development with the most frequent occurrence of IHI (57%) % > ! _^# was that of patients suffering from generalized [23]. We found a high frequency of IHI in this het- cryptogenic epilepsy. Compared to the controls, erogeneous group of patients. In addition, many the occurrence was almost three times higher and ! ' $ reached the same occurrence of IHI as reported that such a high frequency of morphological brain in patients with severe developmental anomalies changes has been detected in epilepsies previously [4,69]. The IHI frequency was high also in focal ! $ cryptogenic epilepsy (4/5 patients), but this sub- has yet to be elucidated. group was very small. Studies including patients with severe develop- Isolated left-sided IHI is far more common than mental brain anomalies present much higher fre- bilateral IHI. This distribution pattern was found quencies of IHI. Sato and coworkers [69] found in- both in controls and in epilepsy patients, except for complete inversion in 64% (28/44) and Baker and those with generalized cryptogenic epileptic syn- Barkovich [4] in 82% of the patients diagnosed dromes where 58% of IHIs were bilateral. Isolated with congenital brain malformations such as cor- right-sided IHI seems to be a very rare entity, and pus callosum agenesis, lissencephaly, or holopros- ! ! ! " encephaly. and in only four epilepsy patients. Laterality has ' not always been mentioned in articles describing such is still unknown. A recent study indicates hippocampal morphological changes including that IHI does not have a detrimental effect on IHI. In literature, isolated rightsided IHI has only memory functioning but is associated with dys- been reported in one patient with partial epilepsy function of other areas [74]. That also supports [7]. In cases with persistent left-sided IHI, cere- the hypothesis that persistent IHI may be a sign bral development may have been disturbed when of a disturbance in brain development with clini- the right hippocampus had completed inversion cal consequences elsewhere in the brain, and new, process but the left side had not yet reached that both clinical and imaging, follow-up studies could developmental stage. be very informative. In TLE patients, there was no statistically signif- IHI cannot be the etiology of epilepsy as such icant difference in frequency, laterality or distri- but could be a sign of general failure in brain de- bution of total and partial forms of IHI compared velopment, which might be responsible, both for with controls. The laterality of the EEG focus did incomplete inversion of the hippocampus and fo- ' $ @ ! ! - cal or multifocal pathological changes in the cer- sality between the focus in TLE and IHI. ebral cortex. Cryptogenic epilepsy refers to a condition for which the nature of the underlying cause is un-
40 Conclusions
CONCLUSIONS
IHI is a common morphologic variation. Leftsided In our fetal MRI study, there were a higher number IHI is more common than bilateral, and isolated of fetuses than in the earlier hippocampal studies right-sided incomplete hippocampal inversion " & ! ! < seems to be very rare. show the wide time variation in the development of the hippocampal region and also that the right side often develops faster. The frequency of IHI is higher in the epilepsy population in comparison to the non-epileptic controls, mainly due to a high proportion of IHI in @ ! " " " generalised cryptogenic epilepsy syndromes. IHI malrotation. In particular, we thought that incom- is not more frequent in TLE than in nonepileptic plete hippocampal inversion is a better descrip- controls and no causality between TLE and IHI tion of this condition because the hippocampus is seems to exist. IHI, a morphological variant of the not “malrotated” but rather probably the inversion hippocampus, is not an aetiological factor in epi- was never completed. lepsy but can be a sign of disturbed cerebral de- velopment that may affect other parts of the brain leading to epilepsy. In our limited case material we cannot prove that the non-ovoid shape precedes the ovoid shape but our data does not contradict that hypothesis. Hippocampal inversion was not completed up to GW 24 in one half of the preterm neonates. From the GW 25 onwards, the frequency of IHI is simi- lar to that in the general adult population, at ap- proximately the 20% level. Hippocampal inver- sion seems to continue longer than presumed on the basis of the earlier studies in limited samples. The higher IHI frequency on the left than right side, demonstrated in earlier studies in children and adults, exists even in preterm neonates.
41 Radiological studies on hippocampal development
SUMMARY IN SWEDISH
Detta avhandlingsarbete avbildar utvecklingen av togenisk epilepsi och Rolandisk epilepsi, men hippocampal regionerna i perioden från 17 - 36 inte bland dem med temporallobsepilepsi. Det var gestationsveckor (GV) med hjälp av prenatal MR- ingen korrelation med epilepsifokus vid EEG och undersökning av foster, samt i perioden mellan 23 IHI. Så trots att förekomsten av IHI var hög i vis- och 35 GV med hjälp av kranialt ultraljud på pre- sa grupper, kunde inte IHI vara orsak till epilepsi matura nyfödda barn. Dessutom har vi genom att utan snarare visat att det någonstans i hjärnan har analysera MR hjärna av 300 epilepsi patienter och skett någon förändring som kunde vara orsak till 150 epilepsifria subjekt, evaluerat förekomst av epilepsi. icke färdigutvecklade hippocampala former och deras relation till olika epilepsisyndromen. I delarbete III har vi studerat normal hippocam- pusutveckling genom att analysera coronara bil- I delarbete I har vi analyserat förekomsten av ”In- der på ultraljud hjärna. Efter att samtliga prematu- complete hippocampal inversion” (IHI) som är " ! % < &% ! ! > en hippocampal formvariant, bland den ”vanliga” analyserades bilder av 158 nyfödda i gestations- populationen, dvs. bland 100 friska frivilliga och ålder 23 till 35 gestationsveckor (GV). Vi har stu- patienter utan hjärnanomalier, tumörer, epilepsi, ! " " " ! " ! hydrocephalus och vilket tillstånd som helst som när hippocampus är fullt utvecklad. Cirka 50% av kunde påverka temporala loben. Vi har studerat prematura har inte fullt utvecklad hippocampus MR hjärna i det coronara planet och funnit att före GV 25. Efter GV 25 var frekvensen av IHI, 19% av de undersökta hade IHI, bilateralt eller på som vi har tolkat som omogen hippocampal form, vänster sida. Samtliga med IHI hade också verti- lika som i barn och vuxna i arbete I och II. kal kollateral sulcus på IHI sida. Detta visar att IHI inte alls är en ovanlig hippocampusform. I manuskriptet under bearbetning som är delarbe- te IV har vi analyserat morfologisk utveckling av I delarbete II har vi analyserat förekomsten av IHI hippocampus under fetalperioden, genom analys hos epilepsipatienter. Kontrollgruppen var 100 av 60 fetala MR-undersökningar och 3 postmor- personer från arbete I plus 50 nya utvalda med tala MR av aborterade foster. Ingen av dem hade samma kriterier. Epilepsigruppen hade i början hjärnpatologi. Vi har analyserat hippocampus- 300 patienter och efter exkluderingen av patien- området under gestationsveckorna 17-36. Det var ter med hjärnanomalier, tumörer, hydrocephalus stora individuella variationer i utvecklingen av och patienter med inkompletta data blev det 201 både hippocampus och collateral sulcus. Ofta sker kvar. 30% av dem hade IHI, och i förhållande till utvecklingen asymmetriskt, och det är höger sida % < ! % % - som utvecklas snabbare. nad. Hög frekvens var bland patienter med kryp-
42 Summary
SUMMARY IN SERBIAN
Ova disertacija odslikava razvoj hipokampusa i na kontrolnu grupu. Visoka ucestalost je nadjena susjednih anatomskih struktura u periodu od 17- kod pacijenata sa kriptogenom epilepsijom kao i 36 gestacione nedjelje koristeci prenatalni pregled kod pacijenata sa rolandickom epilepsijom, ali ne fetusa magnetnom rezonancom (MR), kao i u pe- i medju pacijentima sa epilepsijom temporalnog riodu od 23-35 gestacione nedjelje koristeci ultraz- reznja. Korelacija izmedju epileptickog fokusa na vucni pregled mozga prerano rodjene novorodjen- EEG i NHI nije nadjena. Iako je ucestalost NHI u cadi. Pored toga analizirajuci MR snimke mozga pojedinim grupama visoka NHI se ne moze sma- kod 300 pacijenata koji boluju od epilepsije i 150 trati direktnim uzrocnikom epilepsije, nego je pri- subjekata koji nemaju epilepsiju diskutovali smo je indikator da se tokom razvoja negdje u mozgu ucestalost nepotpuno razvijenih formi hipokam- desila promjena koja je uzrocnik epilepsije. pusa i njihov odnos sa pojedinim epileptickim sin- dromima. Treca studija se bavi normalnim razvojem hipokampusa kroz analizu koronarnih presjeka U prvoj od cetiri studije u ovoj disertaciji ana- ultrazvuka mozga kod 158 prerano rodjene no- lizirali smo ucestalost nekompletne hipokampalne vorodjencadi u starosti od 23 do 35 gestacionih inverzije (NHI) koja predstavlja jednu varijantu nedelja. Posto su iskljucana novorodjencad sa sig- oblika hipokampusa, medju ”normalnom” popu- % " " " * > ! " lacijom, sacinjenom od 100 subjekata, zdravih do- oblik hipokampusa sa ciljem da utvrdimo vrijeme brovoljaca i pacijenata bez anomalija mozga, tu- kada je hipokampus potpuno razvijen. Oko 50% mora, hidrocefalusa, epilepsije ili bilo kog drugog prerano rodjene novorodjencadi nema potpuno patoloskog stanja koje moze uticati na temporalni razvijen hipokampus prije 25 gestacione nedelje, rezanj mozga. Proucavali smo koronarne MR ali je poslije tog perioda procenat novorodjencadi sekvence i nasli da 19% av ispitanika ima zna- sa NHI, hipokampusnom formom koju smatramo kove NHI, obostrano ili na lijevoj strani. Kod svih nezrelom, isti kao kod kontrolne grupe u studiji I koji su imali NHI, vertikalni kolateralni sulkus je i II. takodje postojao na istoj strani na kojoj je i NHI. Cetvrta studija sa manuskriptom u obradi pred- Druga studija obuhvata analizu ucestalosti NHI stavlja analizu morfoloskog razvoja hipokam- kod oboljelih od epilepsije. Kontrolnu grupu su pusa za vrijeme fetalnog perioda. Izucavali smo sacinjavali subjekti iz prvog dijela zajedno sa 50 hipokampalni region kod 60 prenatalnih MR fetu- novih izabranih na osnovu istih kriterija kao za sa i 3 postmortalna MR abortiranih fetusa, gesta- prvu studiju. Grupu oboljelih od epilepsije sacin- cione starosti 17-36 nedelja. Nijedan nije imao pa- javalo je u pocetku 300 pacijenata ali poslije isklju- toloske promjene na mozgu. Uocili smo znacajne cenja iz studije pacijenata sa anomalijama mozga, individualne varijacije u razvoju kako hipokam- tumorima, hidrocefalusom, kao i onih sa nepot- pusa tako i kolateralnog sulkusa. Cesto razvoj ide punim podacima u istoriji bolesti, u studiji je os- asimetricno i pri tome je desna strana ta koja se tao 201 pacijent. 30% od njih je imalo NHI, sto je brze razvija. predstavljalo statisticki znacajnu razliku u odnosu
43 Radiological studies on hippocampal development
44 Acknowledgements
ACKNOWLEDGEMENTS
This thesis is not the work of one person, it has been produced by many sup- porting people and I would like to express my sincere gratitude to everyone who helped me to complete this work:
Professor Raili Raininko, my supervisor, for introducing me to research, all positive advice, patience, constant support and kindness.
My associate supervisors Eva Kumlien and Peter Mattsson for their enthusi- "> ! " $ % ! help in clinical analyze of the patients, kind advices and help in statistical evaluation of the studies.
My co-authors Staffan Lundberg, Chen Wang, Professor Orvar Eeg-Olofsson, Professor Uwe Ewald, and Johan Wikström for generous sharing their knowl- edge and support.
My co-author and Portuguese colleague Nuno Canto Moreira for collecting and generous sharing of his Portuguese patients data.
Håkan Pettersson and Nora Velastegui for their technical support, kindness and patience.
Christl Richter-Frohm for fantastic secretary help.
Peter Pech, Torsten Lönnerholm, Marie Raiend and Eva Penno my colleagues on the Department for paediatric radiology for support during this project and friendship.
Ognjen Gasic for assistance during the patients selection.
All my colleagues and personal at the Department of radiology for helping me in various ways.
45 Radiological studies on hippocampal development
46 References
REFERENCES
1. Almog B, Gamzu R, Achiron R, Fainaru O, Za- 11. Best PJ, White AM. (1999). Placing hippocampal lel Y. (2003). Fetal lateral ventricular width: what single-unit studies in a historical context. Hippo- should be its upper limit? A prospective cohort campus. 9:346-351. study and reanalysis of the current and previous 12. Blackstad TW. (1956). Commissural connections data. J Ultrasound Med. 22:39-43. of the hippocampal region in the rat, with special 2. Arnold SE, Trojanowski QJ. (1996). Human fetal reference to their mode of termination. Journal of hippocampal development: I. Cytoarchitecture, Comparative Neurology. 105:417-528. myeloarchitecture and neuronal morphologic 13. Blackstad T.W. (1958). On the termination of features. J Comp Neurol. 367:274-292. some afferents to the hippocampus and fascia 3. Arnold SE, Trojanowski QJ. (1996). Human fetal dentata. Acta Anatomica 35:202-214. hippocampal development: II. The neuronal cy- 14. Brodal A. (1947). The hippocampus and the toskeleton. J Comp Neurol. 367:293-307. sense of smell. A review. Brain 70:179–222. 4. Baker LL, Barkovich AJ. (1992). The large tem- 15. Brodal A. (1982). Neurological Anatomy. New poral horn: MR analysis in developmental brain York, Oxford Univ. Press. anomalies versus hydrocephalus. AJNR Am J 16. Bronen RA, Cheung G. (1991). MRI of the normal Neuroradiol. 13:115-122. hippocampus. Magn Reson Imaging. 9:497-500. 5. Barry GF, Judith A. (1989). Whitworth Diction- 17. Bronen RA, Cheung G. (1991). MRI of the tem- ary of Medical Eponyms, new edn in 2002. The poral lobe: Normal variations, with special ref- Parthenon Publishing Group. erence toward epilepsy. Magn Reson Imaging. 6. Barsi P, Kenez J, Solymosi D, Kulin A, Halasz 9:501-507. P, Rasonyi G, Janszky J, Kaloczkai A, Barcs G, 18. Bush G, Luu P, Posner MI. (2000). Cognitive and Neuwirth M, Paraicz E, Siegler Z, Morvai M, " \ &$ Jerney J, Kassay M, Altman A. (2000). Hippo- Trends Cogn Sci. 4:215-222. campal malrotation with normal corpus callo- sum: a new entity? Neuroradiology. 42:339-345. 19. Cajal SRY. (1995). Vols I & II Histology of the Nervous System of Man and Vertebrates. Trans- 7. Baulac M, De Grissac N, Hasboun D, Oppen- lated (from the French) by Neely and Larry heim C, Adam C, Arzimanoglou A, Semah F, Swanson. New York, NY: Oxford University Lehericy S, Clemenceau S, Berger B. (1998). Press, Oxford. Hippocampal developmental changes in patients with partial epilepsy: Magnetic resonance imag- 20. Campbell S, Macqueen G. (2004). The role of the ing and clinical aspects. Ann Neurol. 44:223-233. hippocampus in the pathophysiology of major depression. J Psychiatry Neurosci. 29:417-26. 8. Bear M. (2006). Neuroscience exploring the Brain, 3rd edn. Lippincott Williams & Wilkins, 21. Chi JG, Dooling EC, Gilles FH. (1977). Gyral (Chapter I Neuroscience: Past, present and development of the human brain. Ann.neurol. future). 1:86-93. 9. Bechterew W. (1900). Demonstration eines Ge- 22. Clifford RJJ. (1994). MRI based Hippocam- hirns mit Zerstoring der vorderen und inneren pal volume measurement in epilepsy Epilepsia. Teile der Hirnrinde bei der Schläfelappen. Neu- 35:S21-S29. rol Zbl. 19:990-991. $ Q "" Q ! " 10. Bernasconi N, Kinay D, Andermann F, Antel of the International League Against Epilepsy. S, Bernasconi A. (2006). Analysis of shape and $ < ! Q positioning of the hippocampal formation: an Epilepsies and Epileptic syndromes. Epilepsia MRI study in patients with partial epilepsy and 30:389-399. healthy controls. Brain. 128: 2442-2452
47 Radiological studies on hippocampal development
24. Farrell TA, Hertzberg BS, Kliewer MA, Harris 35. Humphrey T. (1967). The development of the hu- L, Paine SS. (1994). Fetal lateral ventricles: reas- " " «$ ^ $ }} ¥$ sessment of normal values for atrial diameter at 36. Isaacs EB, LucasA CWK, Wood SJ, Johnson US. Radiology 193:409-411. CL, Marshall C, Vargha-Khadem F, Gabisn 25. Fatterpekar GM, Naidich TP, Delman BN, Agui- DG. (2000). Hippocampal volume and everyday naldo JG, Gultekin SH, Sherwood CC, Hof PR, memory in children of very low birth weight. Pe- Drayer BP, Fayad ZA. (2002). Cytoarchitecture diatric Res. 47:713–720. of the human cerebral cortex: MR microscopy of 37. Isaacson RL, Kimble DP. (1972). Lesions of the excised specimens at 9.4 Tesla. AJNR Am J Neu- limbic system: Their effects upon hypotheses roradiol. 23:1313-1321. and frustration. Behavioral Biology. 7:767-793. 26. Fisher RS, Van Emde Boas W, Blume W, Elger C, 38. Jackson JH, Beevor CE. (1889). Case of tumour Genton P, Lee P, Engel J Jr. (2005). Epileptic Sei- of the right temporosphenoidal lobe bearing on * ! # ! the localisation of the sense of smell and on the International League Against Epilepsy (ILAE) interpretation of a particular variety of epilepsy and the International Bureau for Epilepsy (IBE). Brain. 12:346-349. Special Article Epilepsia, 46:470–472. 39. Jay V. (2002). Pierre Paul Broca. Arch Pathol Lab 27. Fishman RS. (2007). The Nobel Prize of 1906. Med 126:250-251. Arch Ophthalmol. 125:690-694. 40. Jinkins R. (2000). Atlas of neuroradiologic Em- 28. Gamss RP, Slasky SE, Bello JA, Miller TS, Shin- bryology, Anatomy and variants p. 122 LWW. nar S. (2009). Prevalence of hippocampal mal- 41. Jones EG. (1999). Golgi Cajal and the Neuron rotation in a population whiteout seizure. AJNR Doctrine. Journal of the History of the Neurosci- Am J Neuroradiol. 30:1571-1573. ences. 8:170-178. 29. Garel C, Chantrel E, Elmaleh M, Brisse H, Sebag 42. Jung R, Kornmuller AE. (1938). Eine methodik G. (2003). Fetal MRI: normal gestational land- der ableitung lokalisierter potential schwankin- marks for cerebral biometry, gyration and my- gen aus subcorticalen hirnyebieten. Arch Psy- elination. Childs Nerv Syst. 19:422-425. chiat Neruenkr. 109:1-30. 30. Green, JD, Arduini A. (1954). “Hippocampal ac- 43. Kadar T, Arbel I, Silbermann M, Levy A. (1994). tivity in arousal”. J Neurophysiol 17:533–57. Morphological hippocampal changes during nor- $ \ > < «> «#> \> mal aging and their relation to cognitive deterio- Russell SA, Paley MNJ, Whitby EH. (2010). A ration J Neural Transm Suppl. 44:133-43. prospective study of fetuses with isolated ven- 44. Kesler S, Reiss AL, Vohr B, Watson C, Schnei- triculomegaly investigated by antenatal sonogra- der KC, Katz KH, Maller-Kesselman J, Silbereis phy and in utero MR imaging. AJNR Am J Neu- J, Constable RT, Makuch RW, Ment LR. (2008). roradiol. 31:106-111. Brain volume reductions within multiple cog- 32. Hampel H, Bürger K, Teipel SJ, Bokde AL, nitive systems in male preterm children at age Zetterberg H, Blennow K. (2008). “Core candi- twelve. J Pediatr. 152:513–520. date neurochemical and imaging biomarkers of 45. Kier EL, Fulbright RK, Bronen RA. (1995). Alzheimer’s disease”. Alzheimers Dement. 4: Limbic lobe embryology and anatomy: Dissec- 38–48. tion and MR of the medial surface of the fetal 33. Harrison PJ. (2004). The hippocampus in schizo- cerebral hemisphere. AJNR Am J Neuroradiol. phrenia: a review of the neuropathological evi- 16:1847-1853. dence and its pathophysiological implications. 46. Kier EL, Kim JH, Fulbright RK, Bronen RA. Psychopharmacology (Berl). 174:151-162. (1997). Embryology of the human fetal hippo- 34. Hevner FR, Kinney CH. (1996). Reciprocal en- campus: MR imaging, anatomy, and histology. torhinal-hippocampal connections established by AJNR Am J Neuroradiol. 18:525-532. human fetal midgestation. J Comp Neurol 372: 384-394.
48 References
47. Kruger L. Woolsey TA. (1990). Rafael Lorente 60. Parkinson J. (1817). “An essay on the shaking pal- de Nó: 1902-1990 Obituary. The Journal of Com- sy”. The Journal of Neuropsychiatry and Clinical parative Neurology. 300:1-4. Neurosciences. 2002 (2): 223–236. 48. Kuruba R, Hattiangady B, Shetty AK. (2009). 61. Peltier B, Hurtevent P, Trehan G, Derambure P, Hippocampal neurogenesis and neural stem Pruvo J-P, Soto-Ares G. (2005). IRM des malfor- cells in temporal lobe epilepsy. Epilepsy Behav. mations de l`hippocampe dans l`epilepsie tempo- 1:65-73. rale refractaire. J Radiol. 86:69-75. 49. Kuzniecky R, Jackson GD. (1995). Magnetic 62. Petersen I, Eeg-Olofsson O. (1971) The develop- Resonance in Epilepsy. Raven Press chapter I. ment of the electroencephalogram children from } $ _ ! % \^> _> @ ! ¬> + - the age of 1 through 15 years. Non-paroxysmal radori Tolsa C, Sizonenko S, Lazeyras F, Hüppi activity. Neuropediatrics. 2:247–304. PS. (2008). Intrauterine growth restriction af- 63. Peterson BS, Vohr B, Staib LH, Cannistraci CJ, fects the preterm infant’s hippocampus. Pediatric Dolberg A, Schneider KC, Katz KH, Westerveld Res. 63:438–443. M, Sparrow S, Anderson AW, Duncan CC, 51. Mark LP, Daniels DL, Naidich TP, Hendrix LE. Makuch RW, Gore JC, Ment LR. (2000). Re- (1995) Limbic connections. AJNR Am J Neuro- gional brain volume abnormalities and long-term radiol. 16:1303-1306. cognitive outcome in preterm infants. JAMA 284:1939–1947. 52. Martinussen M, Flanders DW, Fischl B, Busa E, Lø haugen GC, Skranes J, Vangberg TR, Brubakk 64. Prull MW, Gabrieli JDE, Bunge SA. (2000). “Ch AM, Haraldseth O, Dale AM. (2009). Segmental 2. Age-related changes in memory: A cognitive brain volumes and cognitive and perceptual cor- neuroscience perspective”. in Craik FIM, Salt- relates in 15-year-old adolescents with low birth house TA. The handbook of aging and cognition. weight. J Pediatr. 155:848–853. Erlbaum. 53. McEwen BS. (1999). Stress and hippocampal 65. Rados M, Judas M, Kostovic I. (2006). In vitro plasticity. Annu Rev Neurosci. 22:105-122. MRI of brain development. Eur J Radiology. 57: 187-198. 54. Montenegro MA, Kinay D, Cendes F, Bernasconi A, Coan AC, Li LM, Guerreiro MM, Guerreiro 66. Righini A, Zirpoli S, Parazzini C, Bianchini E, CAM, Lopes-Cendes I, Andermann E, Dubeau Scifo P, Sala C, Triulzi F. (2006) Hippocampal F, Andermann F. (2006). Patterns of hippocam- infolding angle changes during brain develop- pal abnormalities in malformations of cortical ment assessed by prenatal MR imaging. AJNR development. J Neurol Neurosurg Psychiatry. Am J Neuroradiol. 27:2093-2097. 77:367-371. 67. Robinson V. (editor). (1939). ”Bell’s Law, within 55. Nakano I. (1998). The limbic system: An outline the description of Bell’s Palsy, including a brief and brief history of its concept. Neuropathology discussion about Charles Bell, 1774-1842”. The Symposium 18:211-214. Modern Home Physician, A New Encyclopedia of Medical Knowledge. WM. H. Wise & Com- 56. Nosarti C, Al-Asady MHS, Frangon S, Stewart pany (New York), page 92. AL, Rifkin L, Murray RM. (2002). Adolescents who were born very preterm have decreased 68. Sasaki M, Sone M, Ehara S, Tamakawa Y. (1993). brain volumes. Brain 125:1616–1623. Hippocampal sulcus remnant: potential cause of change in signal intensity in the hippocampus. 57. Okada Y, Kato T, Iwai K, Iwasaki N, Ohto T, Radiology. 188:743-746. Matsui A. (2003). Evaluation of hippocampal in normal infolding using magnetic resonance im- 69. Sato N, Hatakeyama S, Shimizu N, Hikima A, aging. Neuroreport. 14:1405-1409. Aoki J, Endo K. (2001). MR evaluation of the hippocampus in patients with congenital malfor- 58. O’Keefe J, Nadel L. (1978). The Hippocampus as mations of the brain. AJNR Am J Neuroradiol. a Cognitive Map, Oxford: Oxford U.P. 22:387-936. 59. Papez JW. (1937). A proposed mechanism of emotion. Arch Neurol psychiat. 38:725-743.
49 Radiological studies on hippocampal development
70. Scoville WB, Milner B. (1957). Loss of recent 77. Theysohn JM, Kraff O, Maderwald S, Schla- memory after bilateral hippocampal lesions. mann MU, de Greiff A, Forsting M, Ladd SC, Journal of Neurology, Neurosurgery and Psychi- Ladd ME, Gizewski ER. (2009). The human hip- atry 20:11-21. pocampus at 7 T – In vivo MRI. Hippocampus 71. Small S, Stern Y, Tang M, Mayeux R. (1999). 19:1-7. Selective decline in memory function among ¥$ " ¬> @ ! «> _@> @ ! healthy elderly Neurology. 52: 1392–1396. SK, Lodygensky GA, Anderson PJ, Egan GF, In- 72. Smith DM, Mizumori SJY. (2006). Hippocampal der TE. (2008). Neonate hippocampal volumes: Place Cells, Context and Episodic Memory. Hip- prematurity, perinatal predictors, and 2-year out- pocampus. 16:716-729. come. Ann Neurol. 63:642–651. 73. Starkman NM, Gebarski SS, Berent S, Schtein- 79. Tudor M, Tudor L, Tudor KI. (2005). “[Hans gart DE. (1992). Hippocampal formation volume, Berger (1873-1941) - the history of electroen- memory dysfunction, and cortisol levels in pa- cephalography]”, Acta medica Croatica 59 (4): tients with Cushing’s syndrome. Biological Psy- 307–13. chiatry 32: 756-765 80. Tulving E. (1972). Episodic and semantic mem- 74. Stiers P, Fonteyne A, Wouters H, D’Agostino A, ory. In E. Tulving and W. Donaldson (Eds.), Or- Sunaert S, Lagae L. (2010). Hippocampal malro- ganization of Memory (pp. 381-402). New York: tation in pediatric patients with epilepsy associ- Academic Press. ated with complex prefrontal dysfunction. Epi- 81. Vallejo-Manzur F. (2003). The resuscitation lepsia 51:546-555. greats. Andreas Vesalius, the concept of an arti- ¥}$ " 4 \> ¬ < 4 > @ ! +> + ! < 4 `$ $ } ¥$ (1997). Volume and number of neurons of the hu- 82. Wyldes M, Watkinson M. (2004). Isolated mild man hippocampal formation in normal aging and fetal ventriculomegaly. Arch Dis Child Fetal Alzheimer’s disease. The Journal of Compara- Neonatal Ed. 89:F9–F13. tive Neurology. 379: 482–494. 83. Yakovlev PI. (1948). Motility, behavior and the 76. Talamini LM, Meeter M, Elvevåg B, Murre JM, brain: Stereodynamic organization and neural Goldberg TE. (2005). Reduced parahippocampal coordinates of behavior. J Nerv Ment Dis. 107: connectivity produces schizophrenia-like mem- 313-335. ! " ! - 84. Yi Zhang. (2010). Brain MRI and CT morphol- duced parahippocampal connectivity. Arch Gen ogy in healthy aging and Alzheimer’s disease. Psychiatry. 62:485-493. Thesis for doctoral degree, Karolinska Institutet Stockholm.
50