Visuo-spatial neural response interactions in early cortical processing during a simple reaction time task: a high-density electrical mapping study

Neuropsychologia 39 (2001) 828– 844 www.elsevier.com/locate/neuropsychologia

Visuo-spatial neural response interactions in early cortical processing during a simple reaction time task: a high-density electrical mapping study Micah M. Murray a,b, John J. Foxe a,b,c, Beth A. Higgins a, Daniel C. Javitt a,b,d, Charles E. Schroeder a,b,* a

Cogniti6e Neuroscience and Schizophrenia Program, Cogniti6e Neurophysiology Laboratory, Nathan Kline Institute for Psychiatric Research, 140 Old Orangeburg Road, Orangeburg, NY 10962, USA b Department of Neuroscience, Albert Einstein College of Medicine, 1300 Morris Park A6enue, Bronx, NY 10461, USA c Department of Psychiatry and Beha6ioral Science, Albert Einstein College of Medicine, 1300 Morris Park A6enue, Bronx, NY 10461 USA d Department of Psychiatry, New York Uni6ersity School of Medicine, 550 1st A6enue, New York, NY 10016, USA Received 3 April 2000; received in revised form 3 November 2000; accepted 1 December 2000

Abstract The timecourse and scalp topography of interactions between neural responses to stimuli in different visual quadrants, straddling either the vertical or horizontal meridian, were studied in 15 subjects. Visual evoked potentials (VEPs) were recorded from 64 electrodes during a simple reaction time (RT) task. VEPs to single stimuli displayed in different quadrants were summed (‘sum’) and compared to the VEP response from simultaneous stimulation of the same two quadrants (‘pair’). These responses would be equivalent if the neural responses to the single stimuli were independent. Divergence between the ‘pair’ and ‘sum’ VEPs indicates a neural response interaction. In each visual field, interactions occurred within 72 – 86 ms post-stimulus over parieto-occipital brain regions. Independent of visual quadrant, RTs were faster for stimulus pairs than single stimuli. This replicates the redundant target effect (RTE) observed for bilateral stimulus pairs and generalizes the RTE to unilateral stimulus pairs. Using Miller’s ‘race’ model inequality (Miller J. Divided attention: evidence for coactivation with redundant signals, Cognitive Psychology 1982;14:247–79), we found that probability summation could fully account for the RTE in each visual field. Although measurements from voltage waveforms replicated the observation of earlier peak P1 latencies for the ‘pair’ versus ‘sum’ comparison (Miniussi C, Girelli M, Marzi CA. Neural site of the redundant target effect: electrophysiological evidence. Journal of Cognitive Neuroscience 1998;10:216–30), this did not hold with measurements taken from second derivative (scalp current density) waveforms. Since interaction effects for bilateral stimulus pairs occurred within 86 ms and require interhemispheric transfer, transcallosal volleys must arrive within 86 ms, which is earlier than previously calculated. Interaction effects for bilateral conditions were delayed by :10 ms versus unilateral conditions, consistent with current estimates of interhemispheric transmission time. Interaction effects place an upper limit on the time required for neuronal ensembles to combine inputs from different quadrants of visual space ( : 72 ms for unilateral and : 82 ms for bilateral conditions). © 2001 Elsevier Science Ltd. All rights reserved. Keywords: Redundant target effect; RTE; Event-related potential; ERP; Neurophysiology; EEG

1. Introduction Visual space is perceived as a unitary whole, although its initial cortical representation is divided be* Corresponding author. Tel.: + 1-845-3986539; fax: +1-8453986545. E-mail address: [email protected] (C.E. Schroeder).

tween cerebral hemispheres and further divided between upper and lower banks of the calcarine sulcus. Normal visual perception, therefore, relies on recombination of these anatomically separated representations of visual space. This study used neurophysiological and behavioral methods to examine the mechanisms by which the brain combines inputs from different visual quadrants.

0028-3932/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 8 - 3 9 3 2 ( 0 1 ) 0 0 0 0 4 - 5

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

1.1. Neurophysiological measures of inter-quadrant interactions Interactions between responses to stimulus pairs appearing in different quadrants can be investigated by comparing summed visual evoked potentials (VEPs) to single stimuli presented in isolation to different quadrants with VEPs to the same stimuli, presented simultaneously as pairs. The summed VEP responses from single stimuli in two different quadrants (‘sum’) should be equivalent to the VEP from the same stimuli presented simultaneously (‘pair’) if neural responses to each of the single stimuli are independent. Divergence between ‘sum’ and ‘pair’ VEPs indicates an interaction between the neural responses to the spatially separated stimuli. Several forms of interaction effects have been reported from this comparison. For example, two studies have reported that the ‘sum’ and ‘pair’ VEPs superimpose until between 130 and 150 ms post-stimulus [1,55]. Another study applied this comparison to reveal earlier peak P1 latencies for the ‘pair’ versus ‘sum’ condition [30], but did not report the latency of the interaction effect. However, visual inspection of these data (Fig. 9 of Ref. [30]) would suggest an effect substantially earlier than that reported by either Supek et al. [55] or Ahlfors et al. [1]. Moreover, while interhemispheric interactions have been examined through presentations of bilateral stimulus pairs, intra-hemispheric interactions between upper and lower quadrants, which would follow from presentation of unilateral stimulus pairs, have not been examined. The first goal of the present study was to determine the latency and scalp topography of neural response interactions between stimuli in different quadrants during a simple reaction time task.

1.2. Beha6ioral measures of inter-quadrant interactions A second focus of the current investigation was the reaction time facilitation seen when multiple stimuli are simultaneously presented (e.g. Refs. [18,39]). During a visual simple reaction time (RT) task, wherein subjects make speeded responses to stimulus detection, RTs are faster when stimulus pairs are presented together than when single stimuli are presented in isolation. This RT facilitation has been termed the redundant target effect (RTE) and has been consistently observed in control subjects only under bilateral stimulation conditions [8,27,30,41]. Whether this RTE also occurs in the case of unilateral stimulus pairs (upper and lower quadrants of the same hemifield) is unresolved. Two recent studies — one of patients with callosal pathology [20] and the other of hemispherectomized patients [57] — included a unilateral condition that resulted in an RTE facilitation. Problematic with generalizing this result to intact subjects is that neural reorganization in such patients

829

has yet to be fully described (see Refs. [24,66] for discussion). Moreover, two studies have found that the RTE following bilateral stimulus presentations could be accounted for by different models for patients and intact subjects [8,41]. While the Tomaiuolo et al. study included an experiment with control subjects, there were very few trials per stimulus configuration and the stimuli in the bilateral and unilateral stimulus pair conditions were not at equal eccentricities [57]. To the best of our knowledge, the only other study of control subjects that included a unilateral condition at the same eccentricity as the bilateral condition did not show a consistent within-hemisphere RTE; with only one of the two subjects showing an RTE [41]. We first examined this issue of whether the RTE occurs for unilateral as well as bilateral stimulus pairs located at the same eccentricity from fixation. A related issue concerns the necessity of neural interactions in producing the faster reaction times that define the RTE. Two classes of models of RT data have been proposed to explain the RTE: race models and coactivation models. In race models [39], neural interactions are not required to obtain the RTE. Rather, each stimulus of a pair independently competes for response initiation and the faster of the two mediates the response for any trial. Thus, simple probability summation could produce the RTE, since the likelihood of either of two stimuli yielding a fast reaction time is higher than that from one stimulus alone. In coactivation models [29], neural responses from stimulus pairs interact and are pooled prior to motor response initiation. The threshold for motor response initiation is met earlier for stimulus pairs than for single stimuli. We examined whether or not the race model could fully account for the RTE in each visual field. Violation of the race model would indicate that neural response interactions underlie this RT facilitation, in which case an electrophysiological correlate of the RTE might be expected.

1.3. Relationship between neurophysiological and beha6ioral measures An electrophysiological correlate of the RTE for bilateral stimulus pairs has been proposed [30]. Violation of the race model was associated with earlier peak P1 and N1 component latencies of the VEP recorded to stimulus pairs versus summed responses to single stimuli, apparently tracking the RTE [30]. While this comparison is appropriate for determining the latency of neural response interactions, it is inappropriate for determining if peak P1 latency tracks reaction time. Measurements based on summed data may yield a peak latency that is later than the faster of the two single stimulus responses, because the ‘sum’ VEP reflects contributions from both direct (contralateral) and indirect

830

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

(ipsilateral) brain regions. This can be inferred from the absence of ipsilateral P1 and N1 components in patients with callosal agenesis as well as commisurotomy (reviewed in Ref. [5]). Likewise, previous research with neurologically normal subjects has shown that ipsilateral activity is prolonged by interhemispheric transfer time (IHTT; e.g. Ref. [47]). Peak P1 latency for the ‘sum’ condition will be measured at some time point between peak P1 latencies for the direct and indirect responses. The more stringent analysis is the comparison of the responses from the single stimuli with those from stimulus pairs, directly. While this comparison was made in the Miniussi et al. [30] study, they describe that any amplitude or latency differences may have been the result of volume conduction and base their conclusions solely on the ‘pair’ versus ‘sum’ comparison. In order to minimize confounds of volume conduction, this analysis should be performed on the second spatial derivative of the scalp potential (also termed Laplacian or scalp current density). Indeed, recent research using this approach has revealed no support for a direct relationship between peak P1 latency and simple reaction times to single stimuli [46]. Using this Laplacian approach, we re-examined the relationship between peak P1 latency and reaction times in the context of bilateral and unilateral stimulus pairs. We reasoned that if peak P1 latency is directly related to the RTE, then peak P1 latency in the response to stimulus pairs should also be earlier than that in the response to each single stimulus as well as that from the summed single stimulus responses.

2. Materials and methods

designed to be letter-like in appearance without containing inherent orthographic or semantic information. Single stimuli subtended 0.649 0.18° vertically and 0.409 0.10° horizontally at 155 cm viewing distance. Stimuli appeared either singly or in physically identical pairs, centered 1.7° out from the vertical meridian (eccentricity of 2.40° along a 45° diagonal projection from fixation, giving a 3.35° center-to-center distance between pairs across either the horizontal or vertical meridian). Note that careful consideration must be given to the placement of stimuli in order to ensure that the initial neural representation in cortex is completely lateralized (see Section 4). There were a total of eight presentation positions (Fig. 1b). These positions were the upper left quadrant (ULQ), upper right quadrant (URQ), lower left quadrant (LLQ), lower right quadrant (LRQ), upper field pair (UP), lower field pair (LO), left field pair (LF) and right field pair (RF). Because structured stimuli were used in this study, it is conceivable that subjects could respond to stimulus pairs as if they were one stimulus, despite the relatively large distance between them. The pattern of results, however, would argue against this, since the race model is satisfied (see Section 3). Stimuli appeared for 75 ms duration (850– 2000 ms randomized inter-stimulus interval; ISI) while subjects completed a simple reaction time task. Subjects were instructed to make a button-press response as soon as a stimulus was detected. The entire experiment consisted of at least 20 blocks (mean=449 14), each containing 96 trials. Within any block, each presentation position was equally likely. Thus, there was equal likelihood of either a single stimulus or stimulus pair appearing on any trial. Breaks were encouraged between blocks to maintain high concentration and prevent fatigue.

2.1. Subjects Fifteen (seven female) neurologically normal, paid volunteers, aged 19– 39 years (mean=26.5 9 5.4) participated. All subjects provided written informed consent and the Institutional Review Board of the Nathan Kline Research Institute approved the procedures. All subjects had normal or corrected-to-normal vision and were right-handed (Edinburgh Handedness Inventory; [35]).

2.2. Stimuli and procedure Subjects were presented with ‘falsefont’ stimuli (Fig. 1a) that appeared white on a black background of a computer monitor (Iiyama Vision Master Pro 502, model no. A102GT). ‘Falsefont’ stimuli were used to facilitate the comparison of behavioral results from this study with those from studies using a matching task as well as with planned follow-up studies. Stimuli were

Fig. 1. (a) Samples of stimuli used in this experiment. (b) Illustrations of the presentation positions. Distances and stimulus size are not shown to scale. These positions were the upper left quadrant (ULQ), upper right quadrant (URQ), lower left quadrant (LLQ), lower right quadrant (LRQ), upper field pair (UP), lower field pair (LO), left field pair (LF), and right field pair (RF).

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

831

of just 0.28% of trials within the 100– 600 ms range were faster than 130 ms for any subject. Thus, even allowing for a more conservative range of valid reaction times would not significantly alter our conclusions. The average acceptance rate across both behavioral and EEG artifact rejection criteria was 88% (9 8.2) with a maximum of 97% and minimum of 65% for any stimulus configuration. Epochs of continuous EEG (− 100 ms pre- to 500 ms post-stimulus onset) were averaged from each subject for each of the eight stimulus configurations to compute the visual evoked potential (VEP). Baseline was defined as the epoch from −100 ms to stimulus onset. Under the conditions of this study, responses returned to baseline well within the shortest ISI cutoff (850 ms). The average number of accepted sweeps per condition was 457 (9 132) with the lowest for any subject being 218 and the highest, 688.

2.4. Scalp current density (Laplacian) analysis

Fig. 2. (a) The 64-channel electrode montage, nose reference shown in white, depicted here on a digitized scalp surface. (b) Location and nomenclature of 21 parieto-occipital sites from where scalp current density (SCD) waveforms were derived from spherical spline interpolation of the voltage potentials recorded from the original 64-channel montage shown in (a).

2.3. Data acquisition Reaction time data were obtained through right hand button-presses to a response pad placed along the subject’s midline to minimize spatial stimulus– response compatibility effects [2]. Subjects were instructed to maintain central fixation and perform as quickly and as accurately as possible. Eye position was monitored with horizontal and vertical electro-oculogram (EOG) recordings, with a resolution of 0.5° of visual angle. Along with the task constraints (see Section 4), this helped to control for the possibility that subjects might adopt a bias in eye position that would adversely affect the lateralization of stimuli to one or other hemisphere. Continuous EEG was acquired from 64 scalp electrodes (impedances B5 kV), referenced to nose, bandpass filtered from 0.05 to 100 Hz and digitized at 500 Hz (Fig. 2a). Trials with blinks and eye movements were rejected off-line on the basis of the EOG. An artifact criterion of 960 mV (N =12) or 9 100 mV (N = 3) was used at all other scalp sites to reject trials with excessive EMG or other noise transients. Only trials both meeting the EEG artifact rejection criteria and with reaction times between 100 and 600 ms, in keeping with the ‘irreducible minimum’ [51,67], were included in analyses. Nonetheless, B 2% of trials yielded reaction times beyond this range and an average

Analyses of neural response interactions and of putative electrophysiologic correlates of the RTE are problematic when based on voltage data. For one, the distribution of the scalp potential data is dependent on activity recorded at the reference electrode, which alters the absolute voltage recorded at each other electrode. The second and more serious drawback results from the volume conductive nature of the brain and its coverings, which give rise to widespread dispersion of the volume currents generated in discrete brain regions and in turn, to large-scale superposition of electrical fields at the scalp. The differentiation of each intracranial generator’s contribution to the surface distribution is therefore greatly impaired with the direct analysis of voltage data alone. To address these problems, a common practice when high-density recording montages are used, is to take advantage of the relationship between local current density and field potential defined by Laplace’s equation — so called ‘current-density’ analysis. This mathematical transformation involves calculating the second spatial derivative of the field potential, which is directly proportional to the current density and provides an index of local current flow radial to the scalp surface [37,64]. The small inter-electrode distances of high-density montages optimize this estimation [22]. The SCD technique eliminates the contribution of the reference electrode and mathematically eliminates the voltage gradients due to tangential current flows within the scalp. The major strength of the SCD analysis is the possibility for improved spatial and temporal resolution as a consequence of the reduction in spatial superposition of gradients originating from multiple intracranial sources (see Ref. [46] for demonstration). Nonetheless, it is important to realize that while the SCD technique allows us to resolve activity from underlying neural

832

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

generators with an improved degree of spatial precision, it is still influenced by the dispersion of intracranial volume currents. The effects of superposition of spatially adjacent and temporally overlapping brain electrical sources contribute to ambiguity in interpreting these surface maps. In the present data set, SCD waveforms were computed (see Ref. [13,62] for a similar approach) from spherical spline interpolation of each subject’s surface voltage recordings, according to the methods of Perrin et al. [36] and as implemented in FOCUS (MEGIS Corp.) for 21 parieto-occipital sites (Fig. 2b) modified from the ‘International extended 10– 20 system’ (American EEG Society, 1991). These sites were: P1, P2, P3, P4, P5, P6, Pz, T5, T6, PO1, PO2, PO3, PO4, PO5, PO6, PO7, PO8, O1, O2, Oz and iPOz. Analyses were conducted with this data set of SCD waveforms. Computation of the SCD waveforms for the summed responses from single stimuli was based on the sum of the VEP responses for the single stimuli for each visual field.

3. Results

3.1. Neural interaction effects The group-averaged SCD waveforms derived for 21 parieto-occipital scalp sites indicated that the response to stimulus pairs is of smaller amplitude than the summed responses from single stimuli. In order to determine when neural interactions between the responses to single stimuli reach statistical significance, we contrasted the response from the stimulus pair (‘pair’) and the summed responses from the single stimuli (‘sum’) by calculating point-wise paired t-tests (two-tailed) for each of the 21 group-averaged SCD waveforms. Interaction onset was defined as the first point where a 0.05 a-criterion was satisfied for at least 11 consecutive data points (\20 ms at a 500 Hz digitization rate; see, for example Refs. [13,17,42] for a similar approach). Several studies have used this comparison as an analytical tool [1,13– 15,30,31,55] to determine the latency of neural response interactions. Here, we combined this approach with the SCD mathematical transformation to sharpen the topographic visualization of the data (see Section 2 for details). For left visual field stimuli, an interaction effect began at 74 ms post-stimulus at scalp-site P6. For right visual field stimuli, an interaction effect began at 72 ms at site P5. For upper visual field stimuli, an interaction effect began at 78 ms at site iPOz. For lower visual field stimuli, an interaction effect began at 86 ms at site P5. The SCD waveforms showing the earliest interaction effect are shown in Fig. 3 for each visual field.

In order to register the time course of interactions within the framework of the visual response, we next determined the onset latency of the earliest activity at parieto-occipital scalp sites. For each of the 21 groupaveraged parieto-occipital SCD waveforms, point-wise paired t-tests (two-tailed) were conducted between a zero baseline and each of the ‘pair’ and ‘sum’ conditions for each visual field. The first time point forward from 30 ms post-stimulus onset, where the response exceeded the 0.05 a-criterion for 11 consecutive data points, was labeled as the onset of activity for either condition at that site. We calculated the average onset latency of the ‘pair’ and ‘sum’ conditions for each site and labeled this time point the average onset across the two conditions. For all visual fields, this latency was : 47 ms, consistent with other observations of what has been termed the C1 component [6,13]. For the retinal eccentricities we employed, the C1 component is often of positive polarity and may not be visibly discernible from the P1 component at any given electrode [6,13]. From these onset latencies, we calculated the lag from response onset to earliest interaction effect. These values ranged from 23 to 38 ms. Earliest interaction effects, onset latencies and the lag between these latencies are listed in Table 1 for all visual fields. Sequential SCD topographic maps of the group-averaged ‘sum’ minus ‘pair’ difference are shown in Fig. 4 and illustrate the scalp topography of the interaction effect for each visual field. While the analysis of each SCD waveform provides the latency of the interaction effect, the scalp topography confirms that neural response interactions occur first over parieto-occipital regions, indicating that the interaction effect occurs over neighboring sites. In the cases of the left and right visual fields, response interactions first occur over the contralateral hemisphere, followed by the ipsilateral hemisphere. In the cases of the upper and lower visual fields, response interactions appear over both hemispheres. These difference maps would indicate that the interaction effect for upper field stimulus pairs appears over more inferior scalp sites than that for lower field stimulus pairs.

3.2. Redundant target effect (RTE) 3.2.1. RTE found in each 6isual field Two separate analyses of variance (ANOVA) were conducted to determine if mean reaction times were faster for stimulus pairs versus single stimuli. The first ANOVA tested for an RTE with bilaterally presented stimulus pairs. The within subjects factors were visual field (upper, lower) and stimulus type (left quadrant, right quadrant, pair). The factor of visual field was significant (F(1,14) = 28.79; P B0.0001), with lower field stimuli yielding faster reaction times than upper field stimuli. The factor of stimulus type was also significant

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

(F(2,28) =29.78; P B0.0001), with stimulus pairs yielding faster reaction times than either single stimulus, demonstrating a robust RTE for bilateral stimulus pairs. There was no interaction between the factors of visual field and stimulus type (F(2,28) =0.36; P B0.71). The second ANOVA tested for an RTE with unilaterally presented stimulus pairs. As before, the within subjects factors were visual field (left, right) and stimulus type (upper quadrant, lower quadrant, pair). There was no main effect of visual field (F(1,14) =0.14; PB 0.72). The factor of stimulus type was significant (F(2,28) =90.40; P B0.0001), with stimulus pairs yielding faster reaction times than either single stimulus, thereby demonstrating the RTE for unilateral stimulus pairs. In each visual field (upper, lower, left, and right), follow-up planned comparisons (two-tailed t-tests) revealed significant reaction time facilitation for stimulus

833

pairs versus either of the constituent single stimulus. Mean reaction times for each presentation position are plotted in Fig. 5(a) and the follow-up comparisons are listed in Table 2.

3.2.2. Visual field asymmetries of the RTE The ANOVA testing reaction times from unilateral stimuli also revealed an interaction between the factors of visual field and stimulus type (F(2,28) = 4.19; P = 0.026), indicating an asymmetry in the RTE for left and right visual fields. We therefore calculated the magnitude of the RTE for each visual field (measured as the difference between the reaction time to the stimulus pair and the faster of the two single stimuli) for each subject. The average RTE magnitude was 10.9 ms for left visual field stimuli, 5.9 ms for right visual field stimuli, 4.9 ms for upper visual field stimuli and 6.7 ms

Fig. 3. Group averaged (N =15) SCD waveforms (u= 10 − 6) showing the earliest divergence between the summed responses from single stimuli (‘sum’; blue traces) and the response from the corresponding stimulus pair (‘pair’; black traces) for each visual field. The earliest divergence is indicated with a red line, and the latency is labeled.

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

834

Table 1 Earliest average onset and interaction effect for each visual field Visual field

Onset latency (ms) (site)

Earliest interaction between ‘pair’ and ‘sum’ conditions (ms) (site)

Lag from onset to interaction (ms)

Left Right Upper Lower

47 49 46 48

74 72 78 86

27 23 32 38

(P4) (P3) (iPOz) (PO4)

(P6) (P5) (iPOz) (P5)

for lower visual field stimuli (Fig. 5b). Follow-up comparisons revealed that RTE magnitude in the left visual field was greater than that in the right visual field (t14 =3.416; P B0.005), upper visual field (t14 =4.907; PB 0.001) and lower visual field (t14 =3.743; PB 0.003). None of the remaining comparisons reached statistical significance. However, until the effect of response hand is systematically examined, we are reluctant to overstate any asymmetries in RTE magnitude.

3.2.3. Visual detection does not produce a bilateral field ad6antage We next tested whether or not simple reaction times to stimulus pairs differed across visual fields. Previous studies of choice reaction times during matching tasks show that reaction times are faster for bilateral stimulus pairs (upper or lower visual field) than for unilateral stimulus pairs (left or right visual field). This difference has been termed the bilateral field advantage [25]. No bilateral field advantage was observed during completion of this simple reaction time task (comparison of the black bars in Fig. 5a). Rather, our follow-up comparisons revealed that reaction times to stimulus pairs appearing unilaterally — either the left (248 ms) or right (252 ms) visual fields — were faster than those to stimulus pairs appearing bilaterally in the upper (256 ms) visual field (left versus upper, t14 =4.732, PB 0.001; right versus upper, t14 =4.061, P B0.002). Reaction times to stimulus pairs appearing unilaterally were not significantly different from those to stimulus pairs appearing bilaterally at the lower (250 ms) visual field (left versus lower, t14 =0.992, P = 0.338; right versus lower, t14 =1.971, P =0.069). 3.2.4. Tests of the race model In order to determine if the RTE in any visual field requires a neural interaction explanation, we used Miller’s inequality to distinguish between race and coactivation models [29]. This inequality places an upper limit on the cumulative probability (CP) of a reaction time at a given latency for a stimulus pair. For any latency, t, the race model holds when this CP value is less than or equal to the sum of the CP from each of the single stimuli minus an expression of their joint probability [CP(t)single1 + CP(t)single2 −(CP(t)single1CP(t)single2)]. To

perform this analysis, each subject’s reaction time data for each presentation position were separately treated in the following manner to generate group averaged cumulative probability values. Reaction times within the valid range (100–600 ms) were divided into 21 discrete quantiles from the first to the hundredth percentile in 5% increments (1, 5, 10, 15, . . . , 100%). For each presentation position, the CP at each quantile were group averaged to form an aggregate distribution that preserves the shape of individuals’ data (also termed Vincent averaging [65]). In each visual field, the race model holds over all quantiles (Fig. 5c). Probability summation is therefore sufficient to explain the RTE observed in each visual field.

3.3. Electrophysiological correlate of the RTE A previous study reported that the earlier peak P1 latency from voltage waveforms for stimulus pairs (‘pair’) versus summed responses from single stimuli (‘sum’) represents an electrophysiological correlate of the RTE [30]. We observe a similar pattern in the group averaged voltage waveforms for the ‘pair’ and ‘sum’ conditions for both the upper and lower visual fields (black versus blue traces in Fig. 6a). Mean peak P1 latency was significantly earlier for the ‘pair’ condition than the ‘sum’ condition for both upper field as well as lower field stimuli at scalp sites P3 and P4 (Table 3). However, this pattern does not replicate when measurements are taken from SCD derived waveforms. We found no significant peak latency differences for the ‘pair’ versus ‘sum’ comparison for either the upper or lower field at both derived sites P3 and P4 (black versus blue traces in Fig. 6b). Moreover, if peak P1 latency tracks the RTE, then peak P1 latency for stimulus pairs should also be earlier than peak P1 latency for either single stimulus, mirroring the pattern observed with reaction times. This does not hold. For both the upper and lower visual fields, peak P1 latency for stimulus pairs did not significantly differ from that of the contralateral single stimulus (black versus green and red traces in Fig. 6b). Mean peak P1 latency is plotted in the bar graphs of Fig. 6 and the results of the statistical tests are listed in Table 3.

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

As described in Section 1, VEPs from bilateral stimulus pairs likely represent a convolution of direct projection activity, as well as indirect projection activity resulting from interhemispheric transfer. That is, activity from both hemispheres is volume conducted to the recording site. The unilateral conditions of this study allowed us to investigate the contribution of volume conducted activity from the ‘indirectly-activated’ hemisphere to the electrophysiological correlate of the RTE reported for voltage data by Miniussi et al. [30]. To do this, we group averaged for each condition, the SCD waveform from each subject where P1 amplitude was largest over both direct (contralateral to the stimuli) and indirect (ipsilateral to the stimuli) projection scalp sites, separately. This approach yields an optimized measure of P1 peak latency, given variation in electrode position [19] and cortical geometry both between subjects as well as between hemispheres of the same subject [4,9,52]. A subject’s data were evaluated for inclusion in a given test using the following procedure. For each stimulus condition (pair, sum and each single stimulus condition) and over each hemisphere, we identified the SCD waveform showing maximal P1 amplitude for each subject. In those cases when several neighboring sites yielded large amplitude P1 compo-

835

nents, we selected the site nearest to the center. In some cases, no site yielded a distinguishable P1 component. In these cases, that subject’s data for all stimulus conditions involved in that statistical test were excluded (see Fig. 7). From these measurements over direct projection scalp sites, we find that peak P1 latency for the ‘pair’ condition for both left and right visual field stimuli does not significantly differ from that of the respective ‘sum’ condition or the lower quadrant single stimulus (Fig. 7a,c). In contrast, there is a significant peak latency advantage for the ‘pair’ condition over indirect projection sites, with peak P1 latency for the ‘pair’ earlier than both the ‘sum’ and each single stimulus condition (Fig. 7b,d; Table 4). Thus, for stimulus pairs presented to either the left or right visual field, there is a peak P1 latency advantage over indirect projection scalp sites, mirroring the behavioral results, that is not observed over direct projection sites. However, it should be noted that peak P1 latency over indirect projection sites is : 35–40 ms later than the corresponding value at direct projection sites. Moreover, onset of activity over indirect projection sites occurs at : 80 ms, which follows the neural response interaction effect observed over the direct projection hemisphere.

Fig. 4. SCD topographic maps (u= 10 − 5) of the difference between the summed responses from single stimuli and the response from the corresponding stimulus pair for each visual field. These show the scalp topography of neural response interactions for each visual field. Red isocontour lines represent positive values and blue, negative. Note that the polarity of foci is arbitrary, being dependent upon the direction of the subtraction.

836

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

Fig. 5. (a) Mean (N =15) reaction times (S.E.M. indicated) for stimulus pairs (black bars) and the constituent single stimuli (red and green bars) in each visual field. An RTE is present for each visual field. (b) Mean RTE magnitude for each visual field. RTE magnitude was calculated as the difference between reaction times to stimulus pairs and the more quickly responded to single stimulus on an individual basis. A significant left visual field asymmetry in RTE magnitude is observed. (c) Cumulative probability (CP) distributions for single (light gray solid and dashed traces), pair (black traces), and ‘model’ data (dashed dark gray traces). ‘Model’ data are the output from Miller’s inequality [29]. For each visual field there is no violation of the race model, indicating that probability summation could fully explain the RTE.

3.4. Multiple P1 generators The absence of a relationship between peak P1 latency and the RTE is perhaps unsurprising given the fact that the P1 component represents the activation of multiple neural generators [6,13,16,49,50]. SCD topographic mapping of one subject’s (BB) data in response to a single stimulus presented to the lower right quadrant illustrates this point, revealing three positive foci over the 50– 140 ms post-stimulus epoch (Fig. 8a). Over this epoch, the morphology and distribution of these foci varied, which is indicative of a changing generator configuration. The earliest of these foci lies over the contralateral parieto-occipital scalp. The second focus is also over the contralateral hemisphere, but with a

more lateral occipital distribution. The third focus is over the ipsilateral scalp and has a widespread distribution, spanning over dorsal and lateral scalp sites. The corresponding voltage topography (Fig. 8b) smears the visualization of these multiple generators.

4. Discussion ERP responses to stimulus pairs presented simultaneously in different visual hemifields are not equivalent to the summed ERP responses to the constituent single stimuli presented in isolation. Assuming that stimuli have been appropriately lateralized, the difference between the summed and pair ERPs reflects neural re-

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

sponse interactions following interhemispheric transfer. A similar rationale applies to stimulus pairs presented to upper and lower quadrants of the same visual hemifield. There are two experimental factors that could affect this interpretation: adequacy of fixation and the extent of nasotemporal overlap in the human visual system. Regarding fixation, subjects were instructed to centrally fixate and EOG was monitored with a resolution of 0.5° of retinal angle. While this does not preclude very small eye movements, it does control for deviations in eye position that would be large enough to impair the lateralization of stimuli at 1.7° of retinal eccentricity. In addition, there was no advantage to the subject for a bias or shift in eye position, and in fact, these were discouraged by the nature of the task. Because stimulus appearance was brief, location was unpredictable and rapid response was required, adequate performance on this task required steady central fixation. Regarding nasotemporal overlap, the true extent of this variable is not fully resolved from clinical studies that accurately controlled for eye position. Some have reported foveal splitting [12] and others have provided evidence for : 0.4–0.6° [53,54], 1° [11] or as much as 3.5° [66] of overlap. Because these studies involved patient populations, they complicate the attempt to derive parameters of the normal human visual system. For example, the residual sensitivity to stimuli within an :3.5° regions out from the vertical meridian in two patients with functional hemispherectomy was interpreted to reflect re-routing of retinofugal projections to the intact hemisphere, rather than the extent of nasotemporal overlap, as exemplified in the intact brain [66]. Similarly, Sugishita et al. [54] suggest that the true extent of nasotemporal overlap is 0.6°, but because of uncontrolled eye movements and eccentric fixations by their commisurotomy patients, they felt that it was necessary to place stimuli at no less than 2° eccentricity to ensure lateralized presentation. The suggestion of minimal nasotemporal overlap in humans [11,12,53,54] is consistent with precise functional labeling studies in

837

anesthetized, systemically paralyzed non-human primates using 14C-2-deoxy-D-glucose labeling [61], which find no evidence for nasotemporal retinal overlap.

4.1. Neural interaction effects 4.1.1. Timing The latency of the neural response interactions places an upper limit on the onset of activity in neuronal ensembles responsive to both spatial locations, thus providing a time course for the conjoining of separate visual representations. In the case of unilateral stimulus pairs, brain mechanisms that combine information from the upper and lower quadrants are active within :25 ms of C1 response onset (see Table 1). The difference between interaction onset latencies for unilateral and bilateral stimulus pairs is likely related to interhemispheric transmission time (IHTT; see below). It should be noted that the neural response interactions observed in the present data set occur : 50 ms earlier than in other studies [1,55]. One possible explanation for this difference is that the present study used a simple reaction time task, whereas previous studies were conducted under passive viewing conditions. The question of whether or not the latency of neural response interactions is affected by task parameters will be a topic of future experiments. 4.1.2. A no6el measurement of IHTT The onset latency of the interaction effect is :5 –10 ms longer for bilateral than for unilateral stimulus conditions. This is reasonably attributed to IHTT, and is consistent with some recent estimates [5,21,45,46], but not with others [43,44]. In previous studies, IHTT has been estimated by subtracting the peak latency of a VEP component (either P1 or N1) in response to a laterally presented stimulus, as recorded over the ipsilateral scalp from that over the mirror symmetric site over the contralateral scalp (reviewed in Refs. [5,21]). Our findings highlight limitations on both temporal and spatial accuracy in this approach. First, our data demonstrate that neural interactions evoked by bilat-

Table 2 Mean reaction times for stimulus pairs at each visual field and the comparison (paired t-test, two-tailed) with each of the constituent single stimuli Visual field

RTE?

Stimulus pair mean reaction time (ms)

Single stimulus mean reaction time (ms)

t-value(df); P-value

Left

ã

LF: 248

ULQ: 265 LLQ: 260

t14 =10.672; PB0.001 t14 =12.055; PB0.001

Right

ã

RF: 252

URQ: 264 LRQ: 258

t14 =7.367; PB0.001 t14 =3.846; PB0.001

Upper

ã

UP: 256

ULQ: 265 URQ: 264

t14 =5.075; PB0.001 t14 =6.481; PB0.001

Lower

ã

LO: 250

LLQ: 260 LRQ: 258

t14 =6.947; PB0.001 t14 =7.095; PB0.001

838

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

Fig. 6. Group averaged waveforms for the ‘pair’ (black traces), ‘sum’ (blue traces), as well as both single stimulus (red and green traces) conditions in response to stimuli at the lower visual field. (a) Group averaged (N = 15) voltage waveforms from electrodes P3 and P4. (b) Group averaged SCD waveforms (u=10 − 6) from sites P3 and P4. Bar graphs use the identical color scheme as the waveforms and display mean peak P1 latency for each stimulus condition (S.E.M. indicated).

eral stimulus pairs (which require transcallosal volleys) occur prior to the peak of the P1 component. These early interhemispheric effects would be undetected by measures based on the P1 peak. Second, our data provide evidence that the P1 component receives contributions from multiple generators, with different peak latencies. The measured peak latency will depend on the relative contribution of each of these multiple generators to the surface potential. Thus, IHTT measured by traditional approaches (which sum across all contributions) will preferentially reflect input from the largest amplitude generator, not necessarily the one with the shortest latency. SCD topographic mapping indicated that the interaction effect is first observed over posterior scalp sites (see Fig. 4). A strong implication of these results is that this earlier transcallosal volley is

between brain regions with visual [3,26,32] rather than motoric functions [56].

4.1.3. Locus of 6isual neural response interactions The onset of the transcallosal response indicated by our data is : 72 ms, which would allow :25 ms following the onset of the visual cortical response for intrahemispheric processes prior to activation of callosal transfer in regions with appropriate connectivity. The eccentricity of the stimuli used in this experiment would exclude areas V1 and V2. In V1, callosal afferents are present only at the vertical meridian and extend less than a degree peripherally [7]. Likewise, estimated V2 callosal connectivity is confined to the 2° representation (e.g. Ref. [34] for macaque data and Ref. [7] for human data). Response interactions generated

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

by unilateral stimulus pairs also begin at : 25 ms after the onset of the initial cortical response. In determining a probable locus of these neural response interactions, timing information is only minimally useful because there is widespread activation throughout the visual system by the latency of the interaction effect. Both direct intracranial recordings in monkeys [33,48] and scalp VEP data from humans [13] show activation throughout both the dorsal and ventral stream of the visual system within :30 ms of the initial cortical response. Additionally, these data show a large dorsal stream response latency advantage in both macaques [33,48] and humans [13], which would make dorsal stream regions well-suited for the mediation of the early interaction effect for both unilateral and bilateral stimulus pairs. Studies in the monkey help to restrict candidate regions. The initial stages in the ascending pathway with substantial callosal connections beyond the representation of the vertical meridian include V3A, LIP, MT, V4 and IT [10,23,63]. Hemodynamic [58– 60] and histological [7] studies of corresponding regions in humans corroborate certain of these candidate regions. For example, human area V3A contains a retinotopic representation of the upper and lower quadrants of the contralateral visual field [59], as well as a portion of the ipsilateral visual field [58]. Likewise, human area MT is very coarsely retinotopic [60], contains heavy bands of callosal afferents [7] and responds to ipsilateral visual field stimulation [58]. SCD topographic mapping of the interaction effect epoch revealed that neural response interaction in each visual field first appeared over parieto-occipital scalp regions,

839

consistent with dorsal stream regions. Moreover, the interaction effect showed at least limited retinotopy. In the cases of unilateral stimulation (left and right visual fields), neural response interactions first occurred over the contralateral hemisphere. In the cases of upper and lower field stimulation, there is the suggestion that neural response interactions first occurred over inferior and superior scalp regions, respectively. This would suggest that the interaction effect might first occur in brain regions without full representations of both quadrants.

4.2. Redundant target effect 4.2.1. Retinotopic independence of the RTE We demonstrate an RTE that is independent of the exact combination of adjacent visual quadrants stimulated. While the RTE for bilateral stimulus pairs has been repeatedly demonstrated [8,27,30,41], this is the first demonstration of a robust RTE for stimulus pairs presented unilaterally to different visual quadrants of the same visual hemifield. Previous studies of the RTE either did not include the unilateral condition [8,27,30] or failed to observe a consistent effect, possibly due to methodological factors [41]. 4.2.2. The RTE in relation to the BFA Simple reaction times to unilateral stimulus pairs (left or right visual field) are faster than those to upper field stimulus pairs and not significantly different from those to lower field stimulus pairs. This indicates that simple reaction times are not facilitated by the initial projec-

Table 3 Contrast between comparisons of voltage versus SCD waveforms Waveform type

Visual field and scalp site

Voltage

Upper @P3 @P4 Lower @P3 @P4

SCD

‘Pair’ mean peak P1 latency (ms)

Comparison condition mean peak P1 latency (ms)

t-value(df); P-value

‘Pair’: 118.3 ‘Pair’: 117.5

‘Sum’: 128.5 ‘Sum’: 126.8

t14 =3.100; PB0.008** t14 =4.390; PB0.001**

‘Pair’: 91.5 ‘Pair’: 89.1

‘Sum’: 106.3 ‘Sum’: 104.3

t14 =3.258; PB0.006** t14 =2.245; PB0.041*

Upper @P3

‘Pair’: 105.7

@P4

‘Pair’: 115.9

‘Sum’: 116.5 ULQ: 129.7 URQ: 103.9 ‘Sum’: 117.9 ULQ: 110.9 URQ: 134.8

t14 =1.670; t14 =4.819; t14 =0.292; t14 =0.469; t14 =1.259; t14 =3.528;

P= 0.117 PB0.001** P= 0.774 P= 0.646 P= 0.229 PB0.004**

Lower @P3

‘Pair’: 86.9

@P4

‘Pair’: 84.8

‘Sum’: 93.3 LLQ: 124.4 LRQ: 83.6 ‘Sum’: 87.1 LLQ: 77.6 LRQ: 125.9

t14 =1.570; t14 =6.021; t14 =0.949; t14 =0.403; t14 =1.786; t14 =6.930;

P= 0.139 PB0.001** P= 0.359 P= 0.693 P= 0.096 PB0.001**

840

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

Fig. 7. SCD waveforms (u=10 − 6; black = ‘pair’, red =lower quadrant single stimulus, green =upper quadrant single stimulus and blue = ‘sum’) group averaged across optimal P1 scalp sites over the contralateral (a,c) and ipsilateral (b,d) hemispheres. Bar graphs use the identical color scheme as the waveforms and display mean peak P1 latency for each stimulus condition (S.E.M. indicated).

tion of stimulus pairs to different cerebral hemispheres. The RTE is therefore contrasted with the bilateral field advantage [25], which is the facilitation of choice reaction times for bilateral versus unilateral stimulus pairs. Determining whether this distinction is due to processes of shape discrimination and/or stimulus comparison across spatial locations will require additional experiments.

4.2.3. The RTE in relation to neural response interactions Analyses of the reaction time data with Miller’s inequality revealed that probability summation could

fully account for the RTE in each visual field in the current data set [29]. That is, the stimuli in a pair independently compete for response mediation and the probability of either of the two stimuli yielding a fast reaction time is higher than for a single stimulus. Previous studies of the RTE for bilateral stimulus pairs during a visual simple reaction time task performed by neurologically intact subjects have yielded equivocal results with this analysis. Most have reported conformity to the race model [8,28,41], whereas only one study observed violation of the race model [30]. As mentioned earlier, results from patients are variable, with some patients satisfying and others violating the

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

race model within a study [20] or with patients violating the race model and intact subjects satisfying it [8]. The finding of race model violation by intact subjects [30] may be attributable to group averaging cumulative probability values for the same absolute reaction time, rather than the same quantile across individuals’ data, a practice which obscures individual differences in reaction time distributions (see Ref. [28,40] for discussion). While there is no violation of the race model in our study, neural response interactions nonetheless occur. Obviously, such interactions need not be invoked as a mechanism to explain the RTE in any visual field.

4.3. Electrophysiological correlate of the RTE 4.3.1. The use of 6oltage 6ersus SCD wa6eforms While we were able to replicate the findings of Miniussi et al. [30] when the analysis was performed with voltage waveforms, no such difference was observed between peak latencies of the ‘pair’ and ‘sum’ conditions when the analysis is performed with SCD derived waveforms, which have improved spatiotemporal resolution. Moreover, peak P1 latency for stimulus pairs did not significantly differ from that of the contralateral single stimulus, which would be expected if peak P1 latency truly tracks the RTE and mirrors the pattern of reaction times. For stimuli presented to either the left or right visual field, this pattern also held when comparisons were made from direct, but not indirect, projection scalp sites. The electrophysiological correlate of the RTE described by Miniussi et al. [30] is therefore not substantiated by the current study. 4.3.2. Relations between peak P1 latency and the RTE reconsidered Our findings do suggest a complex relationship between peak P1 latency and reaction time. First, proba-

841

bility summation can fully explain the RTE in each visual field, indicating that neural response interactions are not necessary to account for the behavioral facilitation. Second, although voltage measurements yield a peak P1 latency advantage consistent with the RTE, the comparison of SCD derived waveforms for both upper and lower visual field stimulation revealed no peak P1 latency difference between responses to bilateral stimulus pairs and those to the contralateral (direct projection) single stimulus. Because the contralateral response is the shortest latency single quadrant response, a meaningful ‘pair’ advantage, relative to this response and thus mirroring the behavioral facilitation, would be expected if peak P1 latency tracks the RTE. In the corresponding comparison of responses to unilateral stimulus pairs with those to the constituent single stimuli, there is no peak P1 latency advantage for stimulus pairs over contralateral (direct projection) scalp sites. Interestingly, however, a significant peak P1 latency advantage is observed for stimulus pairs over ipsilateral (indirect projection) scalp sites. The peak P1 latency advantage over the ‘indirect’ hemisphere may reflect amplification due to the neural response interaction that occurred some 30 ms earlier over the ‘direct’ hemisphere.

4.4. Multiple P1 generators The present data indicate that multiple brain regions contribute to the P1 component of the VEP. SCD topographic mapping of the 50–140 ms post-stimulus epoch revealed three foci over the posterior scalp. In agreement, studies of both the monkey [16,48,49] and human [6,13,38,50,55] have shown that multiple brain regions are active during the epoch of the P1 component. Visualization of this complex generator configuration requires the ability to examine data sets from

Table 4 Direct versus indirect optimal scalp sites Visual field and scalp sites

‘Pair’ mean peak P1 latency (ms)

Comparison condition mean peak P1 latency (ms)

t-value(df.); P-value

Left Direct

‘Pair’: 78.9

Indirect

‘Pair’: 118.0

‘Sum’: 88.3 ULQ: 117.7 LLQ: 83.5 ‘Sum’: 136.4 ULQ: 133.5 LLQ: 128.7

t12 =1.647; t12 =5.459; t12 =1.714; t10 =4.300; t10 =2.626; t10 =2.627;

P= 0.126 PB0.0002** P= 0.112 PB0.002** P= 0.025* P= 0.025*

Right Direct

‘Pair’: 86.3

Indirect

‘Pair’: 120.5

‘Sum’: 89.4 URQ: 118.5 LRQ: 84.9 ‘Sum’: 128.4 URQ: 131.1 LRQ: 132.2

t12 =1.563; t12 =5.517; t12 =0.470; t10 =3.677; t10 =2.380; t10 =2.500;

P= 0.144 PB0.00015** P= 0.647 PB0.005** P= 0.039* P= 0.031*

842

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

Fig. 8. (a) SCD topographic maps (u= 10 − 6) of one subject’s (BB) response to a single stimulus shown in the lower right quadrant (LRQ) over the 50– 140 ms post-stimulus epoch. Red isocontour lines represent positive values and blue, negative. Several positive foci are evident, whose morphology and distribution varies over time, indicative of a changing generator configuration and implicating at least three separate intracranial sources. (b) Voltage maps of the same data shown in (a) smear the visualization of these multiple generators.

individual subjects, which in turn requires a large number of trials. Although we failed to find a relationship between peak P1 latency and simple reaction time, the possibility remains that activity in one or more of the brain regions contributing to the P1 component is systematically related to reaction time. Determining which, if any, of the multiple cortical generators of the P1 component are directly related to behavioral measures remains a topic for future research. Future studies will have to account not only for these multiple generators, but also variation in electrode placement [19] and cortical geometry [4,9,52] that may make difficult the correlation of the time series recorded at the scalp with the activity of a particular brain region across a group of subjects.

stimulus pairs. Regarding the mechanism governing the RTE, the present data reveal that the RTE in each visual field can be accounted for by simple probability summation, as described in the race model [39] and neural interactions need not be invoked as a mechanism. We were able to replicate the observation of earlier peak P1 component latencies for the ‘pair’ versus ‘sum’ comparison using voltage waveforms, which has been previously interpreted as an electrophysiological correlate of the RTE [30]. However, the pattern of effects obtained by making this comparison with the more specific SCD measures indicates a more complex interpretation than that suggested by Miniussi et al. [30]. Finally, we extend previous demonstrations that the P1 component is comprised of multiple neural generators.

5. Summary and conclusions Acknowledgements High-density VEP mapping revealed significant interactions between neural responses to stimuli presented simultaneously in separate visual quadrants during a simple reaction time task. These interactions began within :80 ms of stimulus onset and were observed both with unilateral stimulus pairs, straddling the horizontal meridian and with bilateral stimulus pairs, straddling the vertical meridian. Simple reaction times to stimulus pairs, distributed between quadrants either bilaterally or unilaterally, were faster than for a single stimulus in any visual quadrant. This replicates previous demonstrations of the redundant target effect (RTE) for bilateral stimulus pairs and also provides the first clear demonstration of the RTE for unilateral

The authors wish to thank Dr Clifford Saron for insightful discussions and Dr Martin Sliwinski for statistical expertise. Work supported by grants from the NIH (MH49334 and MH01439).

References [1] Ahlfors SA, Simpson GV, Foxe JJ, Hrabe J. Nonlinear spatial interactions in visual cortical responses revealed by high-density recordings of evoked potentials. Society for Neuroscience Abstracts 1997;23:1031. [2] Berlucchi G, Crea F, Di Stefano M, Tassinari G. Influence of spatial stimulus – response compatibility on reaction time of ipsi-

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

[3]

[4] [5]

[6]

[7]

[8] [9]

[10] [11]

[12] [13]

[14]

[15]

[16]

[17] [18]

[19]

[20]

[21]

[22]

[23]

lateral and contralateral hand to lateralized light stimuli. Journal of Experimental Psychology: Human Perception and Performance 1977;3:505 – 17. Braun CMJ, Achim A, Villeneuve L. Topography of averaged electrical activity relating to interhemispheric dynamics in normal humans: where does the critical relay take place? International Journal of Psychophysiology 1999;32:1 –14. Brindley GS. The variability of human striate cortex. Journal of Physiology (London) 1972;25:1 –3. Brown WS, Larson EB, Jeeves MA. Directional asymmetries in interhemispheric transmission time: evidence from visual evoked potentials. Neuropsychologia 1994;32:439 – 48. Clark VP, Fan S, Hillyard SA. Identification of early visual evoked potential generators by retinotopic and topographic analyses. Human Brain Mapping 1995;2:170 – 87. Clarke S, Miklossy J. Occipital cortex in man: organization of callosal connections, related myelo- and cytoarchitecture, and putative boundaries of functional visual areas. Journal of Comparative Neurology 1990;298:188 – 214. Corballis MC. Interhemispheric neural summation in the absence of the corpus callosum. Brain 1998;121:1795 –807. DeYoe EA, Carman GJ, Bandettini P, Glickman S, Wieser J, Cox R, Miller D, Neitz J. Mapping striate and extrastriate visual areas in human cerebral cortex. Proceedings of the National Academy of Sciences USA 1996;93:2382 –6. Felleman DJ, Van Essen DC. Distributed hierarchical processing in the primate cerebral cortex. Cerebral Cortex 1991;1:1 – 47. Fendrich R, Wessinger CM, Gazzaniga MS. Nasotemporal overlap at the retinal vertical meridian: investigations with a callosotomy patient. Neuropsychologia 1996;34(7):637 – 46. Fendrich R, Gazzaniga MS. Evidence of foveal splitting in a commisurotomy patient. Neuropsychologia 1989;27(3):273 – 81. Foxe JJ, Simpson GV. Flow of activation from V1 to frontal cortex in humans: a framework for defining ‘early’ visual processing, Neuroimage 2000; submitted. Foxe JJ, Simpson GV, Ahlfors SP. Parieto-occipital :10 Hz activity reflects anticipatory state of visual attention mechanisms. Neuroreport 2000;9(17):3929 –33. Giard MH, Peronnet F. Auditory-visual integration during multimodal object recognition in humans: a behavioral and electrophysiological study. Journal of Cognitive Neuroscience 1999;11(5):473 – 90. Givre SJ, Schroeder CE, Arezzo JC. Contribution of extrastriate area V4 to the surface-recorded flash VEP in the awake macaque. Vision Research 1994;34:415 – 28. Guthrie D, Buchwald JS. Significance testing of difference potentials. Psychophysiology 1991;28:240 – 4. Hershenson M. Reaction time as a measure of intersensory facilitation. Journal of Experimental Psychology 1962;63:289 – 93. Homan RW, Herman J, Purdy P. Cerebral location of international 10 – 20 system electrode placement. Electroencephalography and Clinical Neurophysiology 1987;66:376 –82. Iacoboni M, Ptito A, Weekes NY, Zaidel E. Parallel visuomotor processing in the split brain: cortico-subcortical interactions. Brain 2000;123(4):759 –69. Ipata A, Girelli M, Miniussi C, Marzi CA. Interhemispheric transfer of visual information in humans: the role of different callosal channels. Archives of Italian Biology 1997;135:169 – 82. Jungho¨ fer M, Elbert T, Leiderer P, Berg P, Rockstroh B. Mapping EEG-potentials on the surface of the brain: A strategy for uncovering cortical sources. Brain Topography 1997;9(3):203 – 17. Kobatake E, Tanaka K. Neuronal selectivities to complex object features in the ventral visual pathway of the macaque cerebral cortex. Journal of Neurophysiology 1994;71:856 –67.

843

[24] Levitan S, Reggia JA. Interhemispheric effects on map organization following simulated cortical lesions. Artificial Intelligence in Medicine 1999;17(1):59 – 85. [25] Ludwig TE, Jeeves MA, Norman WD, DeWitt R. The bilateral field advantage on a letter-matching task. Cortex 1993;29:691 – 713. [26] Marzi CA, Miniussi C, Maravita A, Bertolasi L, Zanette G, Rothwell JC, Sanes S. Transcranial magnetic stimulation selectively impairs interhemispheric transfer of visuo-motor information in humans. Experimental Brain Research 1998;118:435 –8. [27] Marzi CA, Smania N, Martini MC, Gambina G, Tomelleri G, Palamara A, Alessandrini F, Prior M. Implicit redundant-targets effect in visual extinction. Neuropsychologia 1996;34:9 –22. [28] Meijers LMM, Eijkman EGJ. Distributions of simple RT with single and double stimuli. Perception and Psychophysics 1977;22:41 – 8. [29] Miller J. Divided attention: evidence for coactivation with redundant signals. Cognitive Psychology 1982;14:247 – 79. [30] Miniussi C, Girelli M, Marzi CA. Neural site of the redundant target effect: electrophysiological evidence. Journal of Cognitive Neuroscience 1998;10:216 – 30. [31] Molholm S, Foxe JJ, Murray MM, Higgins BA, Schroeder CE, Javitt DC, Ritter W. Multisensory audio-visual interactions in early cortical processing: A high-density electrical mapping and behavioral study. International Multisensory Research Forum, Second Annual Meeting, 2000. [32] Mordkoff JT, Miller J, Roch AC. Absence of coactivation in the motor component: evidence from psychophysiological measures of target detection. Journal of Experimental Psychology: Human Perception and Performance 1996;22:25 – 41. [33] Nowak L, Bullier J. The timing of information transfer in the visual system. In: Kaas JH, Rockland K, Peters A, editors. Cerebral Cortex. New York: Plenum Press, 1998:205 – 41. [34] Olavarria JF, Abel PL. The distribution of callosal connections correlates with the pattern of cytochrome oxidase stripes in visual area V2 of macaque monkeys. Cerebral Cortex 1996;6:631 – 9. [35] Oldfield RC. The assessment and analysis of handedness: The Edinburgh Inventory. Neuropsychologia 1971;9:97 – 113. [36] Perrin F, Pernier J, Bertrand O, Echallier JF. Spherical splines for scalp potential and current density mapping. Electroencephalography and Clinical Neurophysiology 1989;72:184 –7. [37] Perrin F, Bertrand O, Pernier J. Scalp current density: value and estimation from potential data. IEEE Transactions on Biomedical Engineering 1987;34:283 – 8. [38] Portin K, Hari R. Human parieto-occipital visual cortex: lack of retinotopy and foveal magnification. Proceedings of the Royal Society, London: Series B, Biological Sciences 1999;266:981 –5. [39] Raab D. Statistical facilitation of simple reaction times. Transactions of the New York Academy of Sciences 1962;24:574 –90. [40] Ratcliff R. Group reaction time distributions and an analysis of distribution statistics. Psychological Bulletin 1979;86:446 –61. [41] Reuter-Lorenz PA, Nozawa G, Gazzaniga MS, Hughes HC. Fate of neglected targets: a chronometric analysis of redundant target effects in the bisected brain. Journal of Experimental Psychology: Human Perception and Performance 1995;21:211 – 30. [42] Rugg MD, Doyle MC, Wells T. Word and nonword repetition within- and across modality: An event-related potential study. Journal of Cognitive Neuroscience 1995;7:209 – 27. [43] Rugg MD, Lines CR, Milner AD. Further investigation of evoked potentials elicited by lateralized stimuli: Effects of stimulus eccentricity and reference site. Electroencephalography and Clinical Neurophysiology 1985;62:81 – 7. [44] Rugg MD, Lines CR, Milner AD. Visual evoked potentials to lateralized visual stimuli and the measurement of interhemispheric transfer time. Neuropsychologia 1984;22:215 – 25.

844

M.M. Murray et al. / Neuropsychologia 39 (2001) 828–844

[45] Saron CD, Foxe JJ, Simpson GV, Vaughan HG Jr. Inter-hemispheric visuomotor interaction: spatiotemporal electrophysiology related to reaction time. In: Iacoboni M and Zaidel E (Eds.). The Parallel brain: the cognitive neuroscience of the corpus callosum. MIT Press, Cambridge, MA 2001 (in press). [46] Saron CD. Spatiotemporal electrophysiology of intra- and interhemispheric visuomotor integration: relations with behavior. Doctoral Dissertation, Albert Einstein College of Medicine, 1999. [47] Saron CD, Davidson RJ. Visual evoked potential measures of interhemispheric transfer time in humans. Behavioral Neuroscience 1989;103:1115 –38. [48] Schroeder CE, Methta AD, Givre SJ. A spatiotemporal profile of visual system activation revealed by current source density analysis in the awake macaque. Cerebral Cortex 1998;8:575 – 92. [49] Schroeder CE, Steinschneider M, Javitt DC, Tenke CE, Givre SJ, Mehta AD, Simpson GV, Arezzo JC, Vaughan HG Jr. Localization of ERP generators and identification of underlying neural processes. Electroencephalography and Clinical Neurophysiology Supplement 1995;44:55 –75. [50] Simpson GV, Foxe JJ, Vaughan HG Jr, Mehta AD, Schroeder CE. Integration of electrophysiological source analyses, MRI and animal models in the study of visual processing and attention. In: Karmos G, Molnar M, Csepe V, Czigler I, Desmedt JE, editors. Perspectives of Event-Related Potentials. New York: Elsevier, 1995:76 – 92 EEG Supplement, vol. 44. [51] Snodgrass JG, Luce RE, Galanter E. Some experiments on simple and choice reaction time. Journal of Experimental Psychology 1967;75(1):1 –17. [52] Stensaas SS, Eddington DK, Dobelle WH. The topography and variability of the primary visual cortex in man. Journal of Neurosurgery 1974;40:747 –55. [53] Sugishita M, Hamilton CR, Sukama I, Hemmi I. Hemispheric representation of the central retina of commissurotomized subjects. Neuropsychologia 1994;32(4):399 – 415. [54] Sugishita M, Hemmi I, Sakuma I, Beppu H, Shiokawa Y. The problem of macular sparing after unilateral occipital lesions. Journal of Neurology 1993;241(1):1 –9. [55] Supek S, Aine CJ, Ranken D, Best E, Flynn ER, Wood CC. Single vs. paired visual stimulation: superposition of early neuromagnetic responses and retinotopy in extrastriate cortex in humans. Brain Research 1999;830:43 –55.

[56] Thut G, Hauret CA, Morand S, Seeck M, Landis T, Michel C. Evidence for interhemispheric motor-level transfer in a simple reaction time task: an EEG study. Experimental Brain Research 1999;128:256 – 61. [57] Tomaiuolo F, Ptito M, Marzi CA, Paus T, Ptito A. Blindsight in hemispherectomized patients as revealed by spatial summation across the vertical meridian. Brain 1997;120:795 – 803. [58] Tootell RBH, Mendola JD, Hadjikhani NK, Liu AK, Dale AM. The representation of the ipsilateral visual field in human cerebral cortex. Proceedings of the National Academy of Sciences USA 1998;95:818 – 24. [59] Tootell RBH, Mendola JD, Hadjikhani NK, Ledden PJ, Liu AK, Reppas JB, Sereno MI, Dale AM. Functional analysis of V3A and related areas in human visual cortex. Journal of Neuroscience 1997;17:7060 – 78. [60] Tootell RBH, Reppas JB, Kwong KK, Malach R, Born RT, Brady TJ, Rosen BR, Belliveau JW. Functional analysis of human MT and related visual cortical areas using magnetic resonance imaging. Journal of Neuroscience 1995;15:3215 –30. [61] Tootell RB, Switkes E, Silverman MS, Hamilton SL. Functional anatomy of macaque striate cortex II. Retinotopic organization. The Journal of Neuroscience 1988;8(5):1531 – 68. [62] Urbano A, Babiloni C, Onorati P, Babiloni F. Human cortical activity related to unilateral movements. A high resolution EEG study. Neuroreport 1996;8(1):203 – 6. [63] Van Essen DC, Newsome WT, Bixby JL. The pattern of interhemispheric connections and its relationship to extrastriate visual areas in the macaque monkey. Journal of Neuroscience 1982;2:265 – 83. [64] Vaughan HG Jr, Arezzo JC. The neural basis of event-related potentials. In: Picton TW, editor. Human Event-Related Potentials: EEG Handbook. New York: Elsevier, 1988:45 – 95 Revised series, volume 3. [65] Vincent, SB. The function of the vibrissae in the behavior of the white rat, Behavioral Monographs 1912:1. [66] Wessinger CM, Fendrich R, Ptito A, Villemure JG, Gazzaniga MS. Residual vision with awareness in the field contralateral to a partial or complete functional hemispherectomy. Neuropsychologia 1996;34(11):1129 – 37. [67] Woodworth RS, Schlosberg H. Experimental Psychology. New York: Holt, 1954 Revised edition.

.