The Speeding of Voluntary Reaction by a Warning Signal

Psychophysiology, 46 (2009), 225–233. Wiley Periodicals, Inc. Printed in the USA. Copyright r 2009 Society for Psychophysiological Research DOI: 10.1111/j.1469-8986.2008.00716.x

PRESIDENTIAL ADDRESS, 2006 The speeding of voluntary reaction by a warning signal

STEVEN A. HACKLEY Department of Psychological Sciences, University of Missouri-Columbia, Columbia, Missouri, USA

Abstract Warning signals can shorten reaction time (RT) via either a top-down mechanism, temporal attention, or a bottom-up one, phasic arousal. The goal of this review article is to identify the locus at which these processes inﬂuence RT. Electrophysiological and behavioral evidence indicate that the chronometric locus for both modulatory effects lies mainly within a narrow window at the center of the stimulus–response interval. This interval presumably encompasses late perceptual, response selection, and early motor processes. Phasic arousal is theorized to reduce the threshold for response selection within a circuit involving the supramarginal gyrus. A blind-sight study indicates that conscious, cortical level processing is necessary for temporal attention, at least when the warning signal is visual.

Descriptors: Temporal attention, Phasic arousal, ERP/fMRI/EMG, Parkinson’s disease, Blind sight

In the field of attention and performance, warning effects used to temporal orienting is associated with a pattern of brain activation be somewhat dismissively characterized as ‘‘nonselective.’’ that differs substantially from that observed during visuospatial Whereas perceptual resources are selectively focused on ear of orienting. entry in Hillyard’s classical paradigm (Hillyard, Hink, Schwent, The study of temporal orienting began with an experiment & Picton, 1973) or on a particular visuospatial location in performed by the founder of scientific psychology, Wilhelm Posner’s paradigm (Posner, 1980), only diffuse, nonspecific Wundt (1880, pp. 238–239). A simple reaction time (RT) task processes were thought to be engaged by a neutral warning was employed in which the subject released a telegraph key upon signal. This view changed dramatically with the introduction by hearing the impact of a steel ball against a metal plate upon which Coull and Nobre (1998) of a temporal analog of Posner’s it had been dropped. On half of the trials the participant was visuospatial orienting paradigm. The authors argued that allowed to witness the action of the electromagnetic release neutral warning signals do elicit a type of selective attention. It mechanism. This served as the warning signal. As expected, re- is a type of attention in which the individual orients to a particular action times were much shorter on warned than on unwarned point in time. In their revised version of Posner’s well-known task, trials (M 5 125 and 259 ms, respectively). Foreperiod duration a symbolic precue indicates the moment at which an imperative was manipulated by dropping the ball from a height of 5 versus stimulus is likely to occur. Key press latencies are shorter if 25 cm. The warning effect was larger with the longer foreperiod, the imperative does in fact occur at that moment (‘‘valid trials’’) a phenomenon that was later to be investigated systematically by as opposed to at some other point in time (‘‘invalid trials’’). Woodrow (1914). Neuroimaging methods allowed Coull and Nobre to show that Wundt (1896, p. 225) used what we would now call a ‘‘mixture analysis’’ of the RT distribution, coupled with introspective evidence, to argue for separable attention effects involving This article is based on a presidential address given at the 46th annual sensory and motor processes. In his view, subjects can adopt meeting of the Society for Psychophysiological Research. I acknowledge either a sensorial set, in which they wait for a clear perception of my coauthors whose work is reviewed here and the neurological patients the stimulus before responding, or a muscular set, in which who served as participants in the clinical studies. The theoretical devel- they focus on the action of releasing the key. Psychophysiology opment owes a special debt to two previous review papers for which mere now offers abundant evidence that precuing attention citation is not adequate, Niemi and Naä¨ta¨nen (1981) and Jennings and toward characteristics of the stimulus or of the response can van der Molen (2005). Finally, I express special appreciation to Jeff selectively modulate sensory and motor processes, respectively Miller, who was my principal collaborator during the early years of my (e.g., Bonnet, Requıń, & Stelmach, 1991; Mangun, 1995). independent career, and to Fernando Valle-Inclań, who has played this The purpose of the present review is to consider which processes role in more recent years. Address reprint requests to: Steven A. Hackley, Department of Psy- are modulated by a neutral warning signal, that is to say, one that chological Sciences, University of Missouri-Columbia, 210 McAlester precues only the imperative’s time of arrival. More specifically, Hall, Columbia, MO 65211, USA. E-mail: [email protected] the goal is to help explain the RTeffect discovered by Wundt the 225 226 S.A. Hackley year following his creation of the first laboratory of experimental a ‘‘simple systems’’ approach. There are few effects on either RT psychology. or reflex latency as large and robust as those produced by warning, and the basic phenomena have been well conserved across the span of phylogeny. Comparing warning effects on voluntary and reflexive reac- Why Does This Still Matter? tions may be helpful toward understanding temporal orienting because the modulatory patterns are often parallel. For example, Warning effects merit a central role in the field of attention and per- a3 Â 3 manipulation of modalities showed that shortening of formance because of the fundamental nature of space and time. If blink latency and voluntary RT occurs for all combinations of understanding how we orient our attention spatially within the visual, acoustic, and cutaneous warning and response stimuli environment is important, then surely an understanding of tem- (Zeigler, Graham, & Hackley, 2001). Blink latency and volun- poral attention is of similar value. Casting the problem in these tary RT both decrease as a function of foreperiod duration when terms emphasizes the generality of temporal orienting. This type foreperiods of differing lengths are intermixed randomly within a of selective attention may underlie such diverse phenomena as block (e.g., Hackley & Graham, 1987). If the speeding of reflex- Pavlovian conditioning, the Contingent Negative Variation ive eyeblink and voluntary key-press responses obeys similar (CNV), prepulse facilitation, and anticipatory cardiac decelera- laws, then the underlying mechanisms might be the same. For tion, just to mention topics from within psychophysiology. example, if a warning signal facilitates synaptic transmission at Another motivation for studying warning effects is the hope alpha motor neurons or within the earliest segment of the sensory that it will lead to a better understanding of clinical phenomena, pathway, then similar effects of various manipulations should be such as the bradykinesia that characterizes Parkinson’s Disease observed for voluntary and reflexive reactions. (PD). Prior to our work on the topic, several purely behavioral In our Parkinson’s disease study, groups of patients, healthy studies had shown that patients with PD are able to some extent older adults, and college students performed a simple RT task in to benefit from a warning signal, and this benefit varies as a which the imperative stimulus was either an intense white-noise function of foreperiod duration (e.g., Jahanshahi, Brown, & burst or an air puff to the forehead (Jurkowski et al., 2005). Marsden, 1992). However, animal research implied that the loss Participants quickly squeezed a wooden dowel when either of of dopaminergic cells in PD should lead to a reduced effect of these stimuli was received. The warning signal was a visual grat- foreperiod duration (e.g., Brown & Robbins, 1991; MacDonald ing presented randomly 1.5, 4, or 6 s before the imperative. Be- & Meck, 2004). Also, the effect of foreperiod duration is exag- cause the cutaneous and acoustic imperative stimuli were gerated in schizophrenia (Zahn, Rosenthal, & Shakow, 1963), a reflexogenic, it was possible to record eyeblink reflexes in addi- disorder believed to involve excessive sensitivity to dopamine. tion to the voluntary, hand-grip responses. Given the involvement of neuromodulators in mediating warning What we observed was that foreperiod effects on reflexes were effects (discussed below), my colleagues and I thought it impor- quite similar across groups. For voluntary reactions, by contrast, tant to attempt to resolve this discrepancy (Jurkowski, Stepp, & only neurologically normal individuals (younger and older con- Hackley, 2005). To this end, we used electromyographic (EMG) trols) exhibited a variable foreperiod effect. These results con- recordings of voluntary and reflexive reactions to assess forepe- verge with animal data (e.g., Brown & Robbins, 1991) in riod duration effects in individuals whose dopamine system had implying that dopaminergic pathways contribute to temporal been compromised by Parkinson’s disease. attention. Previous studies have identified norepinephrine as the According to a widely held view, random variation of fore- key neuromodulator underlying warning effects (Coull, Jones, period influences temporal attention by means of progressive Egan, Frith, & Maze, 2004; Coull, Nobre, & Frith, 2001; changes in conditional probability. (Los & van den Heuvel, 2001, Stafford & Jacobs, 1990; Witte & Marrocco, 1997). Given the offer an alternative view.) At the shortest interval, the participant role dopamine is known to play in the estimation of time intervals does not know when the imperative will occur, so attention tends (Meck & Benson, 2002), perhaps the deficit we identified in Par- to be distributed more or less evenly across all possible forepe- kinsonian patients indicates an impaired ability to track the pas- riods. As the foreperiod ‘‘ages,’’ the conditional probability of sage of time in the service of response preparation. the imperative increases, with p 5 1 if the next-to-the-last interval At least two neuromodulators contribute to warning effects has passed without presentation of the imperative. Temporal at- and, by the same token, at least two underlying cognitive mech- tention is highly focused on such trials, preparation is maximal, anisms are believed to be involved. Research over the past 40 and reactions are correspondingly fast. A similar explanation years has shown that warning effects are not just due to selective may account for foreperiod effects on brain stem reflexes (e.g., temporal attention. In addition, a brief surge in arousal also Hackley & Graham, 1987). facilitates RT, at least when the foreperiod is brief and the warn- My interest in comparing voluntary and reflexive reactions ing signal is intense (Bertelson & Tisseyre, 1969; Posner, 1978). stems from an article that predates Wundt’s 1880 study. A Vi- Evidence includes the fact that, at short foreperiods, accuracy is ennese physiologist, Sigmund Exner (1874), published a paper in often compromised in favor of speed (Bernstein, Rose, & Ashe, which he compared the latency of voluntary key-press responses 1970), probe H reflexes are potentiated (Scheirs & Brunia, 1985), and reflexive eyeblinks to flashes of light. His stated goal was to and the force of voluntary key-press reactions is sensitive to experimentally investigate the ‘‘simplest mental processes.’’ At warning signal intensity (Ulrich & Mattes, 1996). An effect of the time I came upon this remarkable paper, Eric Kandel was warning signal intensity on RTcan occur even when onset of the doing his Nobel prize-winning work on learning (e.g., Kandel & warning signal follows that of the imperative by as much as 100 Schwartz, 1985). Kandel was able to identify the physical basis of ms (Stahl & Rammsayer, 2005), thereby ruling out temporal memory by studying the modulation of primitive reflexes (i.e., attention. Obviously the word ‘‘warning’’ is a misnomer in such alpha conditioning). It seemed to me that the field of attention contexts. In view of this and of the likelihood that different and performance could similarly benefit from the development of mechanisms predominate at short and long foreperiods, the term Speeding of voluntary reaction by a warning signal 227

‘‘accessory stimulus’’ is preferred to ‘‘warning signal’’ when right or left thumb. At the mirror-image location on the other foreperiods are below 500 ms (Welch & Warren, 1986). side of the display, a single distractor was presented, the Greek letter X. As in Wundt’s (1880) experiment, we investigated temporal attention using a blocked manipulation of foreperiod. A Trisection of RT mild, nonstartling vibration presented to the left and right ankles served as the warning signal. To identify which stages of information processing are speeded On half of the blocks of trials, this tactile warning signal pre- by the phasic arousal and temporal orienting aspects of warning, ceded the imperative by a short, easy-to-estimate interval of 600 my colleagues and I have used lateralized ERP components ms. During the other half, a longer and therefore more difficult- (Gratton, 1998). This is a small but growing class of components to-estimate interval of 3000 ms was used. It was assumed that derived by subtracting the waveform recorded at an electrode temporal attention would be more highly focused during short- ipsilateral to the relevant stimulus, response, or remembered foreperiod trials. Indeed, mean RTwas 83 ms faster during short- event from the corresponding waveform on the contralateral than during long-foreperiod trials. side. The use of subtraction components is critical in temporal The goal of the study was to partition this 83-ms behavioral attention studies because the neural processes of interestFthose effect among the three electrophysiologically defined time seg- occurring between stimulus and responseFoverlap with CNV ments (see Figure 1). In the long foreperiod condition, the av- offset. This large, steep slope varies dramatically across condi- erage time required to locate the target letter, as indicated by tions (e.g., foreperiod duration), making it impossible to accu- N2pc latency, was 208 ms. (This was measured at the 50% am- rately measure the latency of ordinary ERP components that are plitude point of the individual subjects’ waveforms.) The next superimposed upon it. Components that occur subsequent to interval, which included the time needed to decide which hand to CNV resolution, such as the N2 and P3, then overlap with respond with (i.e., the N2pc-to-LRP interval), averaged 160 ms. movement-related potentials. Not surprisingly, studies of tem- After selecting the response, the time that was then required to poral attention that failed to take these difficulties into account press the key averaged 122 ms (i.e., the interval from response- have produced inconsistent results (see Correa, Lupiań˜ez, Ma- locked LRP onset until key press). In the short-foreperiod con- drid, & Tudela, 2006, for an excellent review). dition, these intervals were speeded respectively by 4, 56, and 11 In a recent study, my colleagues and I used two lateralized ms (i.e., by 2%, 35%, and 9%). components, the N2pc and LRP (second prominent negativity at These values are consistent with similar previous electrophys- posterior, contralateral sites; lateralized motor readiness poten- iological findings (for review, see Hackley et al., 2007) if one tial), to localize the effect of foreperiod duration within the S-R considers only the speed of information processing. Prior studies interval (Hackley, Schankin, Wohlschlaeger, & Wascher, 2007). of temporal attention effects on the rapidity of late motor pro- The idea was to use the onsets of these components as temporal cesses have obtained either no effect or only a very small one landmarks to divide the interval from stimulus to response into (e.g., Bausenhart, Rolke, Hackley, & Ulrich, 2006; Mu¨ller- three functionally distinct time segments. The time extending from Gethmann, Ulrich, & Rinkenauer, 2003; Tandonnet, Burle, Vidal, imperative onset until N2pc represents early perceptual activity; & Hasbroucq, 2006). The only previous study of visual processing the time from N2pc until LRP includes late perceptual, decision, that used a subtraction component to avoid contamination by and early motor stages; and the interval extending from LRP CNV offset and motor potentials similarly found no variation in onset until key press is occupied by late motor processes. Which the speed of perceptual analysis (Rudell & Hu, 2001). An exper- one of these time segments is shortened by a warning signal? iment in monkeys using single neuron recordings of stimulus- and These two components have been well characterized anatom- response-related activity found no effect whatsoever on the speed ically and functionally. Magnetoencephalographic recordings of visual processes (Fecteau & Munoz, 2007). Rather, the chro- indicate that the earliest portion of the N2pc originates in the nometric benefit due to warning was localized almost entirely to an superior parietal cortex (BA 7; Talairach coordinates: À 5, À 77, early motor process, one that appeared to be involved in the de- 55), whereas the middle and later parts arise from the posterior cision to respond. Taken as a whole, these data indicate that temporal lobe ( À 33, À 65, À 7; 53, À 45, 1; Hopf, Boelmans, Wundt’s (1880) RT effect was mainly due to the speeding of some Schoenfeld, Luck, & Heinze, 2004). During visual search tasks, process that lies near the middle of the RT interval. this component appears at posterior sites contralateral to the Of course, greater rapidity at one stage, say, deciding which target (Luck & Hillyard, 1994). It signifies that the subject has hand to react with, could be due to a nonchronometric change at identified a probable target and, therefore, has analyzed the vi- an earlier stage, such as higher quality visual analyses. In fact, sual display in significant detail. The time interval from array there is evidence from hemodynamic imaging, electrophysiology, onset until N2pc onset is sensitive to target salience and number and purely behavioral studies that temporal attention can influ- of distractors (Wascher & Wauschkuhn, 1996; Wolber & Wasc- ence processes throughout the S-R interval (e.g., Coull & Nobre, her, 2005). The LRP originates mainly in the hand area of pri- 1998; Ghose & Maunsell, 2002; Hasbroucq et al., 1999; Rolke & mary motor cortex (BA 4: Æ 36, À 12, 60; Leuthold & Jentzsch, Hofmann, 2007). But with regard to the chronometric locus of 2002). The time interval from LRP onset until the key press is Wundt’s effect, it seems clear that the most likely candidates are sensitive to response probability, complexity, and force (Masaki, late perceptual analyses, response selection, and early motor Wild-Wall, Sangals, & Sommer, 2004; Mu¨ller-Gethmann, Rink- processes. enauer, Stahl, & Ulrich, 2000; Smulders, Kok, Kenemans, & Bashore, 1995). The paradigm we used to elicit the N2pc and LRP was a two- Localization within the Middle Interval item visual search task. On one randomly chosen side of the display the Roman letter A or B was presented, for which With regard to the phasic arousal component of warning effects, the subject was required to rapidly execute a button press with the behavioral evidence suggests that response selection is the best 228 S.A. Hackley

interval (Hackley & Valle-Inclań, 1998, 1999; discussed below). A response selection locus is supported by the frequent although not universal finding of speed–accuracy trade-offs when foreperiods are less than 500 ms. Based on this common finding, Posner, Nissen, and Klein (1976) theorized that, ‘‘rather than increasing the discriminability of the imperative stimulus, the accessory causes the subject to respond sooner to the information building up in his memory system. The earlier the response, the less accurate it is’’ (p. 161). To examine accessory effects on the response selection stage in more detail, Fernando Valle-Inclań and I manipulated the diffi- culty of response selection by requiring a choice among either nine or three options (Hackley & Valle-Inclań, 1999). The imperative was the letter T or S presented at fixation in one of five possible colors. The identity of the easily discriminable letter indicated whether the subject should react with the left or right hand. Four of the colors determined which finger should be used and the fifth one meant to withhold the response. One of the participants, for example, reacted to yellow T stimuli with a key press by the left ring finger in the nine-choice condition, used other fingers for the blue, green, and red letters, and withheld her response when the imperative was violet in color. In the three- choice condition, the four Go colors were mapped to a single finger but the No Go color remained the same. Based on previous LRP research (e.g., Hackley & Valle- Inclań, 1998; Miller & Hackley, 1992; Osman, Bashore, Coles, Donchin, & Meyer, 1992), we assumed that subjects could quickly discriminate letter identity and begin preparing the hand of response before perceptual analysis was completed. We further assumed that selection of which finger to use would take con- siderably more time in the nine- than in the three-choice condition (Merkel, 1885; Miller & Ulrich, 1998). Consistent with this assumption, reaction times averaged 391 ms longer in the nine- than in the three-choice trial blocks. The idea underlying the experiment was to contrive a situa- tion in which a large portion of the response selection stage occurred after onset of the LRP. The grand average waveforms shown in the top portion of Figure 2 suggest that this goal was achieved. The absolute onset of the stimulus-locked LRP occurs at the same latency in the three- and nine-choice conditions, reflecting the subject’s decision regarding which hand to use. The slope rises much more slowly in the nine-choice condition, due to the lengthy processing required to discriminate letter color and map that color to a particular finger. On half of the trials, a task-irrelevant tone pip was delivered Figure 1. Lateralized ERP components averaged across 31 participants 83 ms prior to imperative onset. This accessory stimulus led to in a visual search task for short (600 ms) and long (3000 ms) foreperiods. faster reactions (M 5 34 ms; see stimulus-locked EMG wave- a: The N2pc component at parieto-occipital sites (PO7/8) as a function of forms in Figure 2), but at a cost of increased errors. If the portion foreperiod duration. b: Stimulus-locked lateralized readiness potentials of the response selection stage that takes place after LRP onset is (sLRP) at fronto-central electrode sites (FC3/4). In stimulus-locked influenced by phasic arousal, then an effect should have been averaging, the EEG is segmented with respect to stimulus onset; hence, observed either on the amplitude of the LRP or on the time the origin of the abscissa indicates stimulus onset. c: Response-locked, interval extending from LRP onset until key press. As shown in lateralized readiness potentials at sites overlying frontocentral cortex the top, right portion of Figure 2, this was not the case. There was (FC3/4). In response-locked waveforms, 0 on the X axis indicates no measurable effect of accessory stimulation on the response- response onset. Reprinted with permission from Hackley, Schankin, Wohlschlaeger, and Wascher (2007). locked LRP. In the stimulus-locked waveforms the effect averaged 36 ms at the LRP half-amplitude point for both the three- and nine-choice conditions. Evidently the speeding of informa- candidate. We can exclude an early sensory-perceptual locus be- tion processing occurred prior to LRP onset. This effect was then cause accessory stimuli can shorten RTeven when delivered well propagated downstream without further modification to the after onset of the imperative (e.g., Bernstein et al., 1970; Stahl & voluntary response (middle, left graph) and the spontaneous, Rammsayer, 2005). A locus within late motor stages is unlikely postresponse blink (bottom, left graph), which signifies that the because accessory stimuli have no effect on the LRP-to-key-press participant regards his task for that trial as completed. Speeding of voluntary reaction by a warning signal 229

Figure 2. Grand average waveforms for the LRP, voluntary EMG, and spontaneous, postresponse eyeblinks recorded with vertical electrooculograms (vEOG) on Go trials. The left column shows stimulus-locked averages and the right response-locked. Responses were faster when a task-irrelevant accessory stimulus accompanied the visual imperative stimulus (solid lines) as compared to control trials in which the visual stimulus was presented alone (dashed lines). Reprinted with permission from Hackley and Valle-Incla´n (1999).

The conclusion, therefore, is this: If Posner’s (1978) theory is the foot vibration accompanied this visual imperative stimulus. correct and accessory effects are due to a criterion shift resulting Temporal uncertainty was manipulated by means of a clock in speed–accuracy trade-offs, then the locus must be an early count-down procedure (Requıń, Brener, & Ring, 1991). On portion of that process. There is apparently little influence of ‘‘good clock’’ trials, the arrow rotated around its center and then phasic arousal on late portions of response selection or on the changed color when it pointed at the designated location (e.g., speed of downstream motor processes. The next study identified 8:00 for some participants). On ‘‘bad clock’’ trials, by contrast, a possible neuroanatomic locus for the early, modulated portion arrow rotation was unpredictive of imperative onset. Both fac- of response selection. tors (Good/Bad Clock Â Accessory/No Accessory) were manipulated across blocks of trials to permit a blocked (integrated) analysis of the fMRI data. These were supplemented by event- Anatomical Locus of Arousal Effect related analyses linked to clock count-down, imperative onset, and key-press response. Even though the accessory-to-imperative onset asynchrony was Based on the theory developed in the previous section, we very brief in the study just described, 83 ms, it might still have expected that the event-related analysis linked to key-press re- supported some degree of temporal orienting. The goals of the sponses would show that activity within decision-related areas next study (Hackley et al., in press) were to more completely was influenced by the accessory. That is to say, we compared separate phasic arousal from temporal orienting effects and trials with and without foot vibrations during good clock con- identify the associated pattern of brain activity using functional ditions to isolate the effect of phasic arousal on voluntary motor magnetic resonance imaging (fMRI). responses. We expected that decision-related structures such as Eliciting a sudden increase of arousal in the noisy, stressful the supplementary motor area, anterior cingulate, and dorsal confines of an MR scanner is no easy matter. To reduce situa- premotor areas would show differences in activation. However, tional anxiety, we preferentially recruited subjects who had ex- although these three areas were more active during the task than tensive experience participating in MRI studies. As an accessory during rest periods, the planned analyses did not reveal signifi- stimulus, we used an extremely salient although not painful cant variation as a function of accessory stimulation. Of the re- stimulusFan intense vibration to the soles of the participant’s gions that did show effects, the one judged most likely to be bare feet. In the first couple of familiarization trials, the vibration directly involved in response selection was the supramarginal elicited whole-body startle and tickle (foot withdrawal 1 laugh- gyrus (SMG, BA 40; see Figure 3). Activation within or just ter) reflexes. Overt reflexes habituated within three familiariza- above this region was found in a previous study of accessory tion trials, and we know from previous research (Lipp, Kaplan, effects, one that used acoustic accessory stimuli coupled with a & Purkis, 2006) that startle does not contribute to accessory pharmacological manipulation of arousal (Coull, Jones, et al., effects on voluntary RT. 2004). The task was to press a key with the left or right hand de- What we know about the function of the SMG is consistent pending on whether a centrally displayed arrow changed color with the assumption that this structure is sensitive to arousal and from gray to either red or green. On half of the blocks of trials, plays a role in response selection. One line of evidence is that 230 S.A. Hackley

measures of activation separately for the left and right motor cortices. The authors confirmed that the LRP-to-key-press interval is not altered by accessory stimulation. But, in addition, they showed that the accessory generates a nonspecific increase in surface negativity on both sides. Now previous research had shown that an accessory triggers changes in key-press force and reflex amplitude that correlate poorly with RTeffects (e.g., Low, Larson, Burke, & Hackley, 1996; Stahl & Rammsayer, 2005). To account for this, Jepma and colleagues theorized that response force is determined by the activation of the relevant (i.e., con- Figure 3. Left panel: Effects of cutaneous accessory stimulation on tralateral) motor cortex, which is higher on trials with an acces- regional hemodynamic activity under conditions of low temporal sory. By contrast, choice RT is determined by the difference in uncertainty (good clock with accessory minus good clock without accessory). Statistical threshold was set at p 5 .01, k 5 5 voxels in this activation between the relevant and irrelevant motor cortex. This event-related analysis. Unpublished data from Hackley et al. (in press). explanation fits nicely with mathematical models of the choice Right panel: Effects of dexmedotomidine, a norepinephrine a´2 agonist, RT (e.g., Ratcliff & Rouder, 1998), a topic explored by Jepma on the enhancement of activity within Brodmann Area 40 by acoustic and colleagues. accessory stimulation. Reprinted with permission from Coull, Jones, Egan, Frith, and Maze (2004). lesions of this portion of parietal cortex cause apraxia and deficits Can Blind Sight Produce Warning Effects? in haptic perception (Brown, 1972). Functional imaging research shows that SMG activation varies with phasic arousal, as in- The final study to be described also addressed the neuroana- dexed both by skin conductance and phenomenological report tomical locus of warning effects but, in this case, using lesion- (Anders, Lotze, Erb, Grodd, & Birbaumer, 2004). Another behavior analyses. The question was whether it might be possible imaging study suggested that the SMG is involved in the retrieval for a warning signal to exert its effects on RTat subcortical levels. of memory for actions (Russ, Mack, Grama, Lanfermann, & As illustrated by the Parkinson’s experiment described earlier, we Knopf, 2003). In the learning phase of their study, subjects either know that temporal orienting can speed brain stem responses read a verbal description of an action (e.g., ‘‘Cut the bread.’’) or such as the acoustic and cutaneous blink reflexes. We also know else read such a description and then pantomimed the move- that Pavlovian conditioning of the blink reflex does not require ments. Later, as participants lay within the MR scanner, they forebrain circuitry (Thompson, 2005). On the assumption that responded to these descriptions and to foils to indicate which classical conditioning is either a subtype of warning effect or the items they recognized as having previously encountered. The left two concepts are equivalent, it seems possible that subcortical and right supramarginal gyri were more strongly activated for processes might supplement the cortical level mechanisms doc- previously enacted items than for those that had merely been read umented by fMRI and ERP research. during the study phase. To test this possibility, my colleagues and I evaluated the Under the present interpretation, the SMG is the most likely effectiveness of warning signals that were delivered from within locus for phasic arousal effects within the S-R pathway. This the scotomas of patients with cortical blindness (Sonnenberg, arousal somehow lowers the threshold for response as the subject Johnson, Jurkowski, & Hackley, 2006). If responses are faster remembers which key to press for a given arrow color and how to when preceded by an invisible warning signal, this would con- press that key while avoiding head movements that would ob- stitute a type of blind sight (Weiskrantz, 1997) and would support scure the MR scan. the existence of subcortical mechanisms for temporal orienting. The analyses that compared good versus bad clock conditions This was an unusual blind sight experiment in that we used group identified a number of probable loci for modulation within the S- methods rather case studies. Specifically, we tested 13 patients R pathway by temporal orienting. These included a portion of with homonymous hemianopia due to unilateral occipital lobe fusiform gyrus that is involved with attention to color (Coull, damage (usually infarct). The methods were similar to those Vidal, Nazarian, & Macar, 2004) and the hand area of primary employed with the Parkinson’s study described earlier (Jurkow- motor cortex (see also Coull & Nobre, 1998). The latter finding ski et al., 2005), except that a choice rather than simple RT task contrasts with LRP results described above and highlights the was employed. Patients reacted to the white noise bursts and fact that amplitude and latency measures often show a striking forehead air puffs by squeezing the left or right hand grip. Blink degree of independence. Such a dissociation is also evident for reflexes and voluntary reactions were recorded electromyograph- perceptual processes, as illustrated in Figure 1. The latency of ically and were signal averaged, like an ERP. N2pc was not affected by foreperiod duration, but its amplitude Foreperiods varied randomly across trials from among the was. values 80, 120, 200, 400, 600, and 1000 ms. The warning signal A recent study by Jepma, Wagenmakers, Band, and was the brief illumination of one of two gratings. One lamp was Nieuwenhuis (in press) attempted to account for the dissocia- positioned within the patient’s visual field defect and the other at tion between warning effects on latency versus amplitude mea- the mirror image location within the sighted hemifield. Eleven of sures of late motor processes. Their LRP experiment was the participants claimed to have no awareness whatsoever of the modeled after that of Hackley and Valle-Inclań (1998), which blind-hemifield gratings; the other two reported some degree of was similar to the three-choice condition of the experiment de- awareness on a few trials. scribed above (Hackley & Valle-Inclań, 1999). The Dutch study There are several pathways that lead from the retina to brain overcame a limitation inherent in the LRP, that it reflects a stem visual areas and would not have been damaged by the oc- difference between the two hemispheres, by including Laplacian cipital lobe insult. Our assumption was that if brain stem path- Speeding of voluntary reaction by a warning signal 231

the ability of sudden stimuli within the blind hemiﬁeld to unconsciously capture attention (Kentridge, Heywood, & Weisk- rantz, 2004), thereby interfering with responses to the imperative stimuli. Identifying the ability of unconsciously registered visual stimuli to either elicit or beneﬁt from attention will be an important step toward understanding the neural basis of consciousness.

Conclusions

In the history of the study of brain function, some problems were of such obvious importance that they were examined with the first available methods and then again with each new wave of technological advancement. When Wundt founded scientific psychology, mental chronometry was the high-tech methodology of the day. He applied this technology to measure the ability of humans to use temporal orienting to enhance speeded perfor- Figure 4. Grand averaged voluntary EMG responses for trials in which mance (Wundt, 1880). Psychophysiologists have played a central the acoustic imperative stimulus was preceded by a warning signal within role in the study of warning effects from the beginning. For ex- the intact or blind hemifield at an onset asynchrony of 80 versus 1000 ms ample, in William James’ (1890) review of psychophysiological (solid and dashed lines, respectively). Reprinted with permission from research on emotion and motor activity, he included a discussion Sonnenberg, Johnson, Jurkowski, and Hackley (2006). of accessory stimulus effects on both voluntary and reflexive action (Volume II, pp. 379–381). ways can mediate warning effects, we should observe a speeding In the present review, I have attempted to illustrate current of voluntary and reflexive reactions at long relative to short themes in temporal orienting research by showing the relevance foreperiods, as in the Jurkowski et al. (2005) study described of recent data to Wundt’s (1880) RTeffect. These studies support earlier (see also Low et al., 1996; Zeigler et al., 2001). This was several conclusions: not the case. When the warning signal was visible, voluntary reactions were much faster at the longest than at the shortest 1. Although a warning signal can influence processes through- foreperiod, but when it was invisible, there was no effect (Figure out the RT interval, the chronometric effect lies mainly within 4). Similar results were obtained for the blink reflex. These results a narrow window that encompasses late perception, response imply that temporal orienting probably requires cortical-level selection, and early motor processes. processing, at least for visual warning signals. 2. In addition to the well-established role of norepinephrine, the The warning signals in this study were not intended to trigger modulatory effects of dopamine are likely to contribute to phasic arousal. It was necessary to use visual stimuli that were warning effects. small, dim, and had ramped onsets in order to minimize the 3. At short foreperiods, phasic arousal contributes to warning chance of awareness in the blind-hemifield condition. There were effects, perhaps by facilitating the selection of motor mem- some small effects at short foreperiods both for blind- and intact- ories from within the supramarginal gyrus. hemifield warning/accessory signals, but they did not achieve 4. Although purely subcortical mechanisms can mediate certain significance after correction for multiple comparisons and were types of warning effects (e.g., classical conditioning of the not reported. Interestingly, the marginal effect of unseen warn- cutaneous blink reflex), conscious, cortically mediated pro- ing/accessory stimuli on voluntary responses was to lengthen cesses are apparently necessary for temporal orienting initi- rather than shorten RT. If genuine, this finding probably reflects ated by a visual warning signal.

REFERENCES

Anders, S., Lotze, M., Erb, M., Grodd, W., & Birbaumer, N. (2004). Brown, V. J., & Robbins, T. W. (1991). Simple and choice reaction time Brain activity underlying emotional valence and arousal: A response- performance following unilateral striatal dopamine depletion in the related fMRI study. Human Brain Mapping, 23, 200–209. rat. Brain, 114, 513–525. Bausenhart, K., Rolke, B., Hackley, S. A., & Ulrich, R. (2006). The locus Correa, A., Lupiań˜ez, J., Madrid, E., & Tudela, P. (2006). Temporal of temporal preparation effects: Evidence from the psychological attention enhances early visual processing: A review and new evidence refractory period paradigm. Psychonomic Bulletin & Review, 13, 536– from event-related potentials. Brain Research, 1076, 116–128. 542. Coull, J. T., Jones, M. E. P., Egan, T. D., Frith, C. D., & Maze, M. Bernstein, I. H., Rose, R., & Ashe, V. (1970). Preparatory state effects in (2004). Attentional effects of noradrenaline vary with arousal level: intersensory facilitation. Psychonomic Science, 19, 113–114. Selective activation of thalamic pulvinar in humans. NeuroImage, 22, Bertelson, P., & Tisseyre, F. (1969). The time course of preparation: 315–322. Confirmatory results with visual and auditory warning signals. Acta Coull, J. T., & Nobre, A. C. (1998). Where and when to pay attention: Psychologica, 30, 145–154. The neural systems for directing attention to spatial locations and to Bonnet, M., Requıń, J., & Stelmach, G. E. (1991). Changes in electro- time intervals as revealed by both PET and fMRI. Journal of myographic responses to muscle stretch, related to the programming Neuroscience, 18, 7426–7435. of movement parameters. Electroencephalography & Clinical Neuro- Coull, J. T., Nobre, A. C., & Frith, C. D. (2001). The noradrenergic a2 physiology: Electromyography & Motor Control, 81, 135–151. agonist clonidine modulates behavioural and neuroanatomical cor- Brown, J. W. (1972). Aphasia, apraxia, and agnosia: Clinical and relates of human attentional orienting and alerting. Cerebral Cortex, theoretical aspects. Springfield, IL: Charles C Thomas. 11, 73–84. 232 S.A. Hackley

Coull, J. T., Vidal, F., Nazarian, B., & Macar, F. (2004). Functional MacDonald, C. J., & Meck, W. (2004). Systems-level integration of anatomy of the attentional modulation of time estimation. Science, interval timing and reaction time. Neuroscience and Biobehavioral 303, 1506–1508. Reviews, 28, 747–769. Exner, S. (1874). Experimentelle Untersuchung der einfachsten psyc- Mangun, G. R. (1995). Neural mechanisms of visual selective attention. hischen Processe: Zweite Abteilung. [Experimental investigation of Psychophysiology, 32, 4–18. the simplest mental process: Second part.]. Pflueger’s Archiv, 8, 526– Masaki, H., Wild-Wall, N., Sangals, J., & Sommer, W. (2004). The 537. functional locus of the lateralized readiness potential. Psychophysi- Fecteau, J. H., & Munoz, D. P. (2007). Warning signals influence motor ology, 41, 220–230. processing. Journal of Neurophysiology, 97, 1600–1609. Meck, W. H., & Benson, A. M. (2002). Dissecting the brain’s internal Ghose, G. M., & Maunsell, J. H. R. (2002). Attentional modulation in clock: How frontal-striatal circuitry keeps time and shifts attention. visual cortex depends on task timing. Nature, 419, 616–620. Brain and Cognition, 48, 195–211. Gratton, G. (1998). The contralateral organization of visual memory: A Merkel, J. (1885). Die zeitlichen Verha¨ltnisse der Willenstha¨tigkeit [The theoretical concept and research tool. Psychophysiology, 35, 638–647. time relationship of voluntary activity]. Philosophische Studien, 2, 73– Hackley, S. A., & Graham, F. K. (1987). Effects of attending selectively 127. to the spatial position of reflex-eliciting and reflex-modulating stimuli. Miller, J. O., & Hackley, S. A. (1992). Electrophysiological evidence for Journal of Experimental Psychology: Human Perception and Perfor- temporal overlap among contingent mental processes. Journal of Ex- mance, 13, 411–424. perimental Psychology: General, 121, 195–209. Hackley, S. A., Langner, R., Rolke, B., Erb, M., Grodd, W., & Ulrich, Miller, J. O., & Ulrich, R. (1998). Locus of the effect of the number of R. (in press) Separation of phasic arousal and temporal preparation alternative responses: Evidence from the lateralized readiness poten- effects in a speeded reaction time task via fMRI. Psychophysiology. tial. Journal of Experimental Psychology: Human Perception and Per- Hackley, S. A., Schankin, A., Wohlschlaeger, A., & Wascher, E. (2007). formance, 24, 1215–1231. Localization of temporal preparation effects via trisected reaction Mu¨ller-Gethmann, H., Rinkenauer, G., Stahl, J., & Ulrich, R. (2000). time. Psychophysiology, 44, 334–338. Preparation of response force and movement direction: Onset effects Hackley, S. A., & Valle-Inclań, F. (1998). Automatic alerting does not on the lateralized readiness potential. Psychophysiology, 37, 507–514. speed late motoric processes in a reaction time task. Nature, 391, 786– Mu¨ller-Gethmann, H., Ulrich, R., & Rinkenauer, G. (2003). Locus of 788. the effect of temporal preparation: Evidence from the lateralized Hackley, S. A., & Valle-Inclań, F. (1999). Accessory stimulus effects on readiness potential. Psychophysiology, 40, 597–611. response selection: Does arousal speed decision making? Journal of Niemi, P., & Naä¨ta¨nen, R. (1981). Foreperiod and simple reaction time. Cognitive Neuroscience, 11, 321–329. Psychological Bulletin, 89, 133–162. Hasbroucq, T., Osman, A., Possamai, C. A., Burle, B., Carron, S., Depy, Osman, A., Bashore, T. R., Coles, M. G. H., Donchin, E., & Meyer, D. E. D., et al. (1999). Cortico-spinal inhibition reflects time but not event (1992). On the transmission of partial information: Inferences from preparation: Neural mechanisms of preparation dissociated by trans- movement-related brain potentials. Journal of Experimental Psychol- cranial magnetic stimulation. Acta Psychologica, 101, 243–266. ogy: Human Perception and Performance, 18, 217–232. Hillyard, S. A., Hink, R. F., Schwent, V. L., & Picton, T. W. (1973). Posner, M. I. (1978). Chronometric explorations of mind. Hillsdale, NJ: Electrical signs of selective attention in the human brain. Science, 182, Erlbaum. 177–179. Posner, M. I. (1980). Orienting of attention. Quarterly Journal of Ex- Hopf, J. M., Boelmans, K., Schoenfeld, A. M., Luck, S. J., & Heinze, H. J. perimental Psychology, 32, 3–25. (2004). Attention to features precedes attention to locations in visual Posner, M. I., Nissen, M. J., & Klein, R. M. (1976). Visual dominance: search: Evidence from electromagnetic brain responses in humans. An information-processing account of its origins and significance. Journal of Neuroscience, 24, 1822–1832. Psychological Review, 83, 157–171. Jahanshahi, M., Brown, R. G., & Marsden, C. D. (1992). Simple and Ratcliff, R., & Rouder, J. N. (1998). Modeling response times for two- choice reaction time and the use of advanced information for motor choice decisions. Psychological Science, 9, 347–356. preparation in Parkinson’s disease. Brain, 115, 539–564. Requıń, J., Brener, J., & Ring, C. (1991). Preparation for action. In J. R. James, W. (1890). Principles of psychology. New York: Holt & Co. Jennings & M. G. H. Coles (Eds.), Handbook of cognitive psycho- Jennings, J. R., & van der Molen, M. W. (2005). Preparatory for speeded physiology: Central and autonomic nervous system approaches (pp. action as a psychophysiological concept. Psychological Bulletin, 131, 357–448). New York: John Wiley. 434–459. Rolke, B., & Hofmann, P. (2007). Temporal uncertainty degrades per- Jepma, M., Wagenmakers, E.-J., Band, G., & Nieuwenhuis, S. (in press). ceptual processing. Psychonomic Bulletin and Review, 14, 522–526. The effects of accessory stimuli on information processing: Evidence Rudell, A. P., & Hu, B. (2001). Does a warning signal accelerate the from electrophysiology and diffusion-model analysis. Journal of Cog- processing of sensory information? Evidence from recognition po- nitive Neuroscience. tential responses to high and low frequency words. International Jurkowski, A. J., Stepp, E., & Hackley, S. A. (2005). Variable foreperiod Journal of Psychophysiology 41 deficits in Parkinson’s disease: Dissociation across reflexive and vol- , , 31–42. Russ, M. O., Mack, W., Grama, C. R., Lanfermann, H., & Knopf, M. untary behaviors. Brain and Cognition, 58, 49–61. Kandel, E. R., & Schwartz, J. H. (1985). Principles of neural science. (2003). Enactment effect in memory: Evidence concerning the func- Amsterdam: Elsevier. tion of the supramarginal gyrus. Experimental Brain Research, 149, Kentridge, R. W., Heywood, C. A., & Weiskrantz, L. (2004). Effects of 497–504. temporal cueing on residual visual discrimination in blindsight. Scheirs, J. G. M., & Brunia, C. H. M. (1985). Achilles tendon and surface Neuropsychologia, 37, 479–483. EMG activity during anticipation of a significant event and prepa- Leuthold, H., & Jentzsch, I. (2002). Distinguishing neural sources of ration for a voluntary movement. Journal of Motor Behavior, 17, 96– movement preparation and execution: An electrophysiological anal- 109. ysis. Biological Psychology, 60, 173–198. Smulders, F. T. Y., Kok, A., Kenemans, J. L., & Bashore, T. R. (1995). Lipp, O. V., Kaplan, D. M., & Purkis, H. M. (2006). Reaction time The temporal selectivity of additive factors effects on reaction process facilitation by acoustic task-irrelevant stimuli is not related to startle. revealed in ERP component latencies. Acta Psychologica, 90, 97–109. Neuroscience Letters, 409, 124–127. Sonnenberg, D., Johnson, L. N., Jurkowski, A., & Hackley, S. A. (2006). Los, S. A., & van den Heuvel, C. E. (2001). Intentional and unintentional Facilitation as well as inhibition of the blink reflex by a visual prepulse contributions to nonspecific preparation during the foreperiod. requires intact striate cortex. Clinical Neurophysiology, 117, 2284– Journal of Experimental Psychology: Human Perception and Perfor- 2291. mance, 27, 370–386. Stafford, I. L., & Jacobs, B. L. (1990). Noradrenergic modulation of the Low, K. A., Larson, S. L., Burke, J., & Hackley, S. A. (1996). Alerting masseteric reflex in behaving cats. II. Physiologic studies. Journal of effects on choice reaction time and the photic eyeblink reflex. Neuroscience, 10, 99–107. Electroencephalography and Clinical Neurophysiology, 98, 385–393. Stahl, J., & Rammsayer, T. H. (2005). Accessory stimulation in the time Luck, S. J., & Hillyard, S. A. (1994). Spatial filtering during visual search: course of visuomotor information processing: Stimulus intensity Evidence from human electrophysiology. Journal of Experimental effects on reaction time and response force. Acta Psychologica, 120, Psychology: Human Perception & Performance, 20, 1000–1014. 1–18. Speeding of voluntary reaction by a warning signal 233

Tandonnet, C., Burle, B., Vidal, F., & Hasbroucq, T. (2006). Knowing Wolber, M., & Wascher, E. (2005). The posterior contralateral negativity when to respond and the efﬁciency of the cortical motor command: A as a temporal indicator of visuo-spatial processing. Journal of Psy- Laplacian ERP study. Brain Research, 1109, 158–163. chophysiology, 19, 182–194. Thompson, R. F. (2005). In search of memory traces. Annual Review of Woodrow, H. (1914). The measurement of attention. Psychological Psychology, 56, 1–23. Monographs, 17, 1–158. Ulrich, R., & Mattes, S. (1996). Does immediate arousal enhance Wundt, W. (1880). Grundzu¨ge der physiologischen Psychologie ([Foun- response force in simple reaction time tasks? The Quarterly Journal dations of physiological psychology] Leipzig, Germany: Engelmann. of Experimental Psychology, 49A, 972–990. Wundt, W. (1896). Outlines of psychology. Leipzig, Germany: Engelm- Wascher, E., & Wauschkuhn, B. (1996). The interaction of stimulus- and ann. response-related processes measured by event-related lateralizations Zahn, T. P., Rosenthal, D., & Shakow, D. (1963). Effects of irregular of the EEG. Electroencephalography and Clinical Neurophysiology, preparatory intervals on reaction time in schizophrenia. Journal of 99, 149–162. Abnormal and Social Psychology, 67, 44–52. Weiskrantz, L. (1997). Consciousness lost and found: A neuropsychological Zeigler, B. L., Graham, F. K., & Hackley, S. A. (2001). Cross-modal exploration. Oxford, UK: Oxford University Press. warning effects on reﬂexive and voluntary reactions. Psychophysiol- Welch, R. B., & Warren, D. H. (1986). Intersensory interactions. In K. R. ogy, 38, 903–911. Boff, L. Kaufman, & J. P. Thomas (Eds.), Handbook of perception and human performance (vol 1, pp. 25.1–25.36). New York: Wiley. Witte, E. A., & Marrocco, R. T. (1997). Alteration of brain noradrenergic activity in rhesus monkeys affects the alerting component of covert orienting. Psychopharmacology, 132, 315–323. (Received August 20, 2007; Accepted April 1, 2008)