Bilateral field advantage and evoked potential interhemispheric transmission in commissurotomy and callosal agenesis

Neuropsychologia 37 (1999) 1165±1180

www.elsevier.com/locate/neuropsychologia

Bilateral ®eld advantage and evoked potential interhemispheric transmission in commissurotomy and callosal agenesis Warren S. Brown a, b,*, Malcolm A. Jeeves c, Rosalind Dietrich d, Debra S. Burnison a a

The Travis Institute for Biopsychosocial Research, Fuller Graduate School of Psychology, Pasadena, CA, USA Department of Psychiatry and Biobehavioral Sciences, Brain Research Institute, University of California, Los Angeles, CA, USA c Department of Psychology, University of St Andrews, St Andrews, UK d Department of Neuroradiology, University of California, Irvine, CA, USA

b

Received 19 May 1998; accepted 9 December 1998

Abstract The role of the corpus callosum versus other cerebral commissures in the interhemispheric integration of visual information was studied in four individuals with complete agenesis of the corpus callosum, two individuals with partial agenesis, one total commissurotomy patient, and normal individuals. Evoked potential (EP) indices of interhemispheric transmission of visual sensory responses were observed during matching of unilateral and bilateral visual ®eld letters and patterns. Neither the commissurotomy nor any of the acallosal patients had ipsilateral hemisphere visual EPs (P1 and N1), demonstrating that the posterior callosum is necessary for interhemispheric transmission of these components of visual evoked potentials. While the commissurotomy patient could not compare bilaterally presented letters, the anterior commissure of the acallosal patients appeared to be sucient for interhemispheric comparison of single letters. However, bilateral comparison of more complex visual patterns resulted in considerable diculty for complete agenesis patients, while comparison of patterns was more nearly normal when anterior callosal ®bers were present (partial agenesis). # 1999 Published by Elsevier Science Ltd. All rights reserved.

1. Introduction Although a large portion of the posterior corpus callosum is made up of ®bers crossing between the extrastriate visual processing areas of the two hemispheres, the speci®c role of the corpus callosum in the integration of information from the two visual ®elds is as yet uncertain. Numerous reports have appeared of behavioral studies of visual cross integration in commissurotomized patients. The majority of experiments with commissurotomy patients suggest that surgical cutting of all the cerebral commissures leaves the patient unable to integrate or compare bilaterally presented visual stimulus information [73,75±78]. Nevertheless, a few studies have reported that splitbrain subjects may be able to accurately compare * Corresponding author. Tel.: +818 584 5525; fax: +818 584 9630.

simple, easily encoded properties of visual stimuli across the midline [34,50,70]. However these results remain controversial [15] since it is not clear whether information might have been cross-signaled by the subjects in some bodily form. Studies of individuals with total agenesis of the corpus callosum (acallosals) provide additional evidence for the role of the corpus callosum in bilateral integration of visual information [37]. This research generally indicates that acallosal subjects can perform bilateral comparisons of visual stimuli in experimental tasks in which a commissurotomy patient would show no evidence of interhemispheric transfer [27,31,36,38,49,67]. However, this same research shows that the performance of acallosal subjects on interhemispheric comparisons of bilaterally presented visual information may be in some circumstances de®cient compared to normals [27,38]. These data suggest that,

0028-3932/99/$ - see front matter # 1999 Published by Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 8 - 3 9 3 2 ( 9 9 ) 0 0 0 1 1 - 1

1166

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

whereas the corpus callosum is not necessary for comparing certain stimuli presented to the two visual ®elds, the callosum contributes to accuracy and eciency. An important dimension mediating the success of interhemispheric comparisons of information may be the complexity of the information to be transferred [27,38]. An example of the importance of complexity in the adequacy of callosal transfer can be found in studies of interhand transfer of tactile learning in the Tactile Performance Test (Form Board). The standard administration of this test is to have subjects put all of the blocks in the appropriate locations on the board (while blindfolded) using the dominant hand, then to do the same with the nondominant hand on a second attempt at the same board. Normal individuals show considerable saving when shifting to the nondominant hand. Early reports had suggested that individuals with complete agenesis of the corpus callosum have diculty transferring the tactile-spatial learning between hands [22,29,30,57,63,71]. However, Sauerwein and her colleagues reported no de®cits in interhand transfer of tactile memories in a test using the 6-block form of the puzzle. A subsequent study using an 8-block version of the TPT [23] indicated mild de®cits in acallosals. Finally, Sauerwein and Lassonde [68] reported a marked de®cit in interhand transfer on the TPT using the same ACC subjects previously tested, only this time using the 12-block version. Thus, while interhand transfer of simple tactilespatial information (6-block patterns) was possible in acallosals, increasingly signi®cant de®cits were apparent as the complexity of the information to be transferred increased from 6 to 8 to 10-block patterns. It is possible that the same is true of the interhemispheric transfer of visual information. In normal individuals, Jeeves and Lamb [32] demonstrated an e€ect on interhemispheric transmission rate of complexity de®ned as the amount of distortion in bilaterally presented visual patterns that subjects were asked to compare. 1.1. Bilateral ®eld advantage The bilateral ®eld advantage (BFA) is an index of the speed and accuracy in comparing two visual stimuli when presented simultaneously one stimulus to each visual ®eld, compared to performance when both stimuli are presented to the same visual ®eld. Speci®cally, there is an advantage in both reaction time and the percentage of correct trials when two visual stimuli to be compared are presented bilaterally (one to each visual ®eld) rather than unilaterally [3,5,6,11,18,19,32,40,42]. There is some evidence that in normals this advantage for bilateral trials increases as the diculty of the task increases [3,48,51]. Because the stimuli in the two hemispheres must be compared

before a decision is made, it is reasonable to assume that the BFA relies in some manner on callosally mediated interhemispheric interactions. It is clear that commissurotomy patients would not be able to make the bilateral matches on the speci®c tasks in which normals manifest a bilateral advantage [15,16,34,77,78]. However, to the degree that the corpus callosum (vs other cerebral commissures or subcortical pathways) is critical to the BFA, the bilateral advantage should diminish or disappear in acallosal patients, suggesting a decreased ability to compare bilaterally presented visual information. 1.2. Evoked potential indices of interhemispheric transfer An increasing number of studies have used visual evoked potentials (EPs) as an electrophysiological probe of the transfer of stimulus-locked neural activity between the hemispheres in normals [10,11,12,41,59± 61,65], acallosals [10,62], and commissurotomy patients [10,43]. In normals, the early P1 (approximately 100 ms post stimulus) and the N1 (approximately 150 ms) components of the cross-callosal evoked potential are consistently found to be smaller in amplitude and delayed in latency in comparison to the response recorded over the directly stimulated hemisphere. The latency delay (between approximately 10 and 15 ms) has consistently been interpreted as an index of interhemispheric transfer time (IHTT) (see Brown et al. [12] for a meta-analysis of EP-IHTT). This EP method has been demonstrated to be a reliable method of measuring IHTT [65]. Converging evidence suggests that EP-IHTT is sensitive to the development and integrity of the corpus callosum. Salamy [64] studied somatosensory EPs and found that EP-IHTT was signi®cantly negatively correlated with age in the 3.75±13 year range, but not signi®cantly correlated in the 10±20 year range (when the callosum is already mature). Similarly, Thompson et al. [79] have recently reported that both visual EP interhemispheric transfer time and the BFA manifest a developmental progression that reaches asymtote between 6 and 12 years of age, suggestive of the developmental progression of myelinization of the corpus callosum [2,80]. Thus, EP-IHTT is suciently sensitive to re¯ect developmental changes in the corpus callosum. Rugg et al. [60] have presented data demonstrating that the cross-callosal P1 and N1 is absent in individuals with agenesis of the corpus callosum, despite normal appearing waveforms over the directly stimulated hemisphere. Similarly, one abstract has been published suggesting the possibility of a similar EP pattern in commissurotomy patients [43]. To our knowledge, replications of these studies have not been reported,

25

19

15

18

27

40

AC

JD

MM

BE

MO

LB

M

M

M

M

M

M

F

Sex

R

R

L

R

L

L

R

Hand

Split-Brain

Complete ACC

Complete ACC

Complete ACC

Complete ACC

Partial ACC

Partial ACC

Callosum

No

Yes

Yes

Yes

Yes

Yes

Yes

Ant. Comm.

ACC=agenesis of the corpus callosum; na=not available.

14

KF

1

Age

ID

Table 1 Description of acallosal patients1

115 (na)

91 (87,99)

97 (100/95)

108 (97/114)

87 (80/96)

90 (74/111)

80 (78/85)

WAIS/WISC FSIQ (VIQ/PIQ) Reading 4 percentile Spelling 1 percentile Math 1 percentile Reading 25 percentile Spelling 8 percentile Math 21 percentile Reading 75 percentile Spelling 37 percentile Math 5 percentile Reading 23 percentile Spelling 1 percentile Math 5 percentile Reading 81 percentile Spelling 50 percentile Math 39 percentile Reading 39 percentile Spelling 27 percentile Math 7 percentile (na)

WRAT

Ritalin

Obsessive-Compulsive Heterotopia of the left frontal grey matter

Complete commissurotomy for epilepsy

Left parasagittal interhemispheric cyst

Cyst left of the falx cerebri

Dilantin

Dilantin for seizure control

Depakote

None

Zoloft

Arachnoid cyst in 3rd ventricle

Enlarged lateral ventricles

None

Medications

None Behavioral discontrol

Other Neuropathology or Psychopathology

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180 1167

1168

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

nor have the cross-callosal visual EPs of agenesis and commissurotomy patients been studied using the same methods in the same laboratory. What is more, EP patterns of individuals with partial callosal agenesis have not been reported.

1.3. EPs and bilateral visual matching in callosal absence The opportunity was therefore taken to apply the same techniques both behaviourally (unilateral vs bilateral letter or pattern matching) and electrophysiologically in the same laboratory to the study of visual interhemispheric processing in four complete callosal agenesis patients, two patients with partial agenesis, and one surgical commissurotomy patient. The brains of commissurotomy patients and acallosals di€er in many ways but a total commissurotomized patient and a complete agenesis patient share the absence of the linking of the two cerebral hemispheres through the major neocortical commissure, the corpus callosum. Since the surgical patient reported here had all of his cerebral commissures cut, whereas the acallosals we tested were known to have the anterior commissure present, it was of special interest in such a comparative study to examine the possible role of the anterior commissure in the interhemispheric integration of visual information. By similar logic, the relative contribution of anterior vs posterior callosal pathways to bilateral visual integration can be surmised by di€erences between complete and partial agenesis of the corpus callosum. Since the corpus callosum develops from anterior to posterior, partial agenesis typically involves absence of some portion of the posterior callosum [55]. In the study reported here, we sought to compare the abilities of individuals with complete callosal agenesis (but veri®able anterior commissures), individuals with partial callosal agenesis, and a total commissurotomy patient. The focus of this study is the speed and accuracy of comparisons of two stimuli that initially arrive separately in each cerebral hemisphere (i.e., a paradigm that consistently elicits a BFA in normals). The two stimuli are presented on di€erent trials either both in the same visual ®eld, or one stimulus in each visual ®eld, with the subject required to respond as to whether the stimuli match. Simultaneously recorded EPs provide, in the case of the unilateral presentations, an electrophysiological index of interhemispheric visual sensory transfer. 0

ACC=agenesis of the corpus callosum; na=not available.

2. Methods 2.1. Subjects Four individuals with total agenesis of the corpus callosum (ACC), two individuals with partial callosal agenesis, and one commissurotomy patient served as experimental subjects (Table 1). Data from these patients were compared to those of 20 right-handed normal adults previously described in Brown and Jeeves [11], and to data from 15 older normal adolescent boys. Complete agenesis of the corpus callosum patients were males ranging in age from 15 to 27 years, including two left handers and two right handers. Callosal absence was diagnosed by MRI, and, in all cases, the anterior commissure could be visualized. Any other abnormalities seen on MRIs are noted in Table 1. Full scale IQs ranged from 87 to 108. Reading scores on the Wide Range Achievement Test (WRAT) ranged from the 81 percentile to the 23 percentile, while scores on math were markedly lower in all cases. MRIs of the two individuals with partial ACC reveal the presence of only the genu in patient AC, and genu plus a very small part of the anterior midbody in KF, with the anterior commissure present in both. These two patients di€ered from each other in age (teenager vs young adult), gender, and handedness. KF had borderline FSIQ and uniformly low WRAT scores, although it is suspected that some diculty in perseverance in test taking may have contributed to these low scores. AC had more normal range FSIQ and WRAT scores, with little di€erence between reading and math. The commissurotomy patient (LB) was one of the California series of patients [7,8,9,73±76]. MRI revealed complete transection of the corpus callosum, but resolution of the images obtained did not permit conclusive statements regarding the anterior commissure [8]. LB was tested at the age of 40 years. His IQ before surgery was 115; not changing appreciably after surgery. Data from these patients were compared to that of two groups of normal control subjects. First, data for both behavioral performance and EPs were compared to the data from normal right-handed young adults run on the same experiments in the same laboratory, previously reported by Brown and Jeeves [11]. This group consisted of 20 right-handed individuals (9 males and 11 females) between 17 and 50 years of age (mean age=28.9 25.8 years). Second, data for visual matching performance were compared to a new group of 15 normal right-handed older adolescent boys between 15 and 18 years of age (mean age=14.72 1.0 years). The adult group represents the best comparison for the results of patients AC, MO and LB, while the

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

1169

Fig. 1. Evoked potentials for individual patients MM (complete agenesis), JD (complete agenesis), and LB (total commissurotomy) compared to the grand average EPs of 20 normal adults (Nrm). The left column of traces were recorded over the left parietal-temporal cortex, and the right column recorded over the right parietal-temporal cortex. Solid lines are responses to unilateral LVF stimuli, and dashed lines from unilateral RVF stimuli. On the time scale at the bottom the vertical marker indicates onset of the visual stimulus (stim), the height of this bar indicating 5 microvolts of amplitude.

adolescent group is the most appropriate comparison group for KF, JD, MM, and BE. Control subjects reported no history of signi®cant neurological disease, neurosurgery, or serious traumatic head injury, and vision was normal or corrected to normal. Handedness in all normals and patients was con®rmed by the Edinburgh Handedness Inventory [52]. 2.2. Procedure 2.2.1. Letter matching Subjects participated in a letter matching task in which two letters were presented on each trial and subjects asked to determine if the letters constituted a `match' or a `no-match' [11,12,35,42,54]. Upper and lower case `A' and `B' composed the set of possible stimulus letters, with a `match' being considered a letter with the `same name' regardless of whether it was upper or lower case. (e.g., `A' and `a'). Two stimulus letters were presented on each trial in two of four possible letter locations forming the corners of a rectangle surrounding a constantly visible ®x-

ation point (Fig. 1 in Brown and Jeeves, [11]). Letter locations were 28, 19 ' of visual angle to the right and left of the ®xation, and 18, 56' above and below the ®xation. At the viewing distance, stimulus letters were approximately 27 ' of visual angle tall. Six types of trials were presented: both letters in the left visual ®eld (LVF), both letters in the right visual ®eld (RVF), bilateral presentations at upper or lower horizontal positions, or bilateral diagonal presentations in either upper right-to-lower left or upper left-to-lower right directions. Stimuli were presented on a computer CRT (standard character size) for a 60 ms exposure duration. A warning tone preceded stimulus onset by 500 ms. Intertrial intervals varied between 1.5 and 2.0 s. `Match' and `no-match' responses were made by pressing the `N' and `M' keys of a computer keyboard. Subjects alternated between pressing with the right hand (index ®nger `N' and middle ®nger `M'), or left hand (index ®nger `M' and middle ®nger `N'). Stimuli were presented in blocks of 48 trials containing 12 LVF, 12 RVF, 12 bilateral horizontal and 12 bilateral diagonal presentations, with an equal number of

1170

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

`match' vs `no-match', and `same' vs `di€erent case' within each trial type. For each trial, both the reaction time and the response outcome (correct or incorrect), along with information on the type of trial were stored in a computer ®le for later analysis. For either the reaction times or response accuracy (expressed in percent of trials resulting in an error), single-trial data were analysed for each commissurotomy or agenesis patient by ANOVA (unilateral vs bilateral stimulus presentations). Data for controls were analysed by betweensubject ANOVAs of each subject's across-trial mean for the same task dimensions. Comparisons of normals and patients proceeded on the basis of the pattern of performance abilities and de®cits for the patients with respect to the two normal groups. 2.2.2. Pattern matching Patients and the 15 adolescent boys were also run on a version of a bilateral and unilateral patternmatching task previously reported by Larson and Brown [35]. Subjects were instructed to watch the screen for presentations of two patterns of six small circles, and to report with the keyboard response whether the patterns matched or not. Patterns consisted of small circles (the letter `o') appearing in six locations of a 3  3 matrix of potential locations (Fig. 1 in Larson and Brown, [35]). Seven di€erent patterns were used, appearing at the same visual angles horizontally and vertically with reference to the ®xation point as the letter stimuli. Patterns were somewhat larger than the letter stimuli (approximately 18 30 ' high and 18 wide). All of these patterns di€ered from one another in form. There were no cases of the same form di€ering merely in up-down or right-left reversals of orientation. As with the letter task, `match' and `no-match' responses were made by pressing the `N' and `M' keys and responding hand was alternated. Stimuli were presented in blocks of 48 trials containing 12 LVF, 12 RVF, 12 bilateral horizontal and 12 bilateral diagonal presentations, with an equal number of `match' and `no-match' trials within each stimulus con®guration. Response accuracy and reaction times were analysed in the same manner as the letter data. 2.3. Evoked potential recording While subjects performed the letter detection task, evoked EEG potentials (EPs) were recorded synchronized to the onset of the visual presentation. Although EPs were not recorded in the normal adolescent group, Thompson et al. [79] have recently demonstrated adult-like waveforms and EP-IHTTs in adolescents. For EP recording, electrodes were attached to right (RtP/T) and left (LtP/T) parietal/temporal loci (mid-

way between Pz and the interaural line over each hemisphere), referenced to linked earlobe electrodes. For patient KF, three additional electrodes were placed at the parietal midline (Pz), and midway between the parietal/temporal loci and the midline over each hemisphere. In all subjects, electrodes were placed above and at the outer canthus of the right eye for bipolar recording of eye movements. Subjects were grounded at Cz. EEG records began 140 ms prior to stimulus onset and extended 500 ms post-stimulus (2 ms per sample). EEGs were ampli®ed 20,000 and ®ltered 0.1 to 100 Hz (3db down). EPs at each recording electrode from each single trial were stored on computer disk for later averaging. Files were marked as to whether data from this trial exceeded an artifact rejection criterion (i.e., exceeded a preset amplitude threshold). Trials containing artifact were not included in subsequent averages. For each electrode, EPs to correct trials were averaged separately for the two unilateral presentation conditions. Averages were also computed for the 2  2 combinations of left vs right hand responding, and LVF vs RVF. Since EPs partitioned according to responding hand did not suggest di€erences when over plotted, this dimension was not further analysed. Averaged EPs were displayed on computer CRT and the P1 and N1 components identi®ed in the waveforms recorded over the right and left hemispheres for LVF and RVF stimulation. The P1 and N1 components are the primary visual evoked potentials components related to the occurrence of a visual stimulus, but not dependent on further cognitive processing. EPs of the various commissurotomy or agenesis patients were individually compared to the grand mean (acrosssubject) average EPs of normals on the basis of qualitative di€erences in the appearance (presence or absence) of the directly projecting (contralateral) versus the cross-callosal (ipsilateral) P1 and N1 components. 3. Results 3.1. Evoked potentials The normal pattern of EPs recorded over the right and left posterior parietal area in response to the LVF and RVF visual presentations in this task are illustrated in the `Nrm' responses of Fig. 1 [11]. A clearly visible P1/N1 complex is present in the direct-path, contralateral waveforms (dashed line for left parietal and solid line for right parietal). A cross-callosal P1/ N1 component is also clearly visible (solid line for left parietal and dashed line for right parietal), although delayed in latency due to interhemispheric transmission time. These early components (P1 and N1) are fol-

Fig. 2. Evoked potential traces for partial agenesis patient KF. Traces are from a coronal row of recording loci extending from left parietal-temporal (LPT) to left parietal (LP) to midline parietal (unlabeled) to right parietal (RP) to right parietal-temporal (RPT). Response to unilateral LVF stimuli are represented in dashed lines, and unilateral RVF stimuli in solid lines. The inset box is a graph of the amplitude of the N1 potential (peaking at approximately 200 ms) at each recording site, measured against the average amplitude at this time point for all ®ve recordings. Solid and dashed lines represent responses to unilateral RVF and LVF stimuli, respectively.

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180 1171

1172

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

Table 2 Mean error rate and reaction time for bilateral and unilateral presentationsa % Error

A. Letters Adults (Nm) Boys (Nm) KF (Ap) AC (Ap) JD (Ac) MM (Ac) BE (Ac) MO (Ac) LB (SB) B. Patterns Boys (Nm) KF (Ap) AC (Ap) JD (Ac) MM (Ac) BE (Ac) MO (Ac) a

Reaction time

Trials (N)

Unilateral

Bilateral

432 336 432 432 576 856 480 384 1192

16.6 23.8 15.0 4.8 10.0 30.4 35.4 25.1 32.6

11.9 18.2 5.5 4.5 10.0 26.3 17.4 21.9 48.2

336 228 384 528 288 384 384

23.7 10.5 8.0 24.5 22.9 22.4 24.0

18.3 9.9 16.6 29.8 40.1 32.2 23.2

Unilateral

Bilateral

P < 0.001 P < 0.008 P < 0.002 ns ns P = 0.182 P < 0.001 ns P < 0.001 (D)

897 859 1029 962 606 885 841 1327 1104

809 816 886 889 591 821 893 1330 1149

P < 0.001 P < 0.003 P < 0.001 P < 0.006 P = 0.190 P < 0.005 ns ns P < 0.037 (D)

P < 0.012 ns P < 0.02 (D) P = 0.176 (D) P < 0.002 (D) P < 0.03 (D) ns

776 1087 931 546 715 775 1147

751 1003 975 540 759 748 1176

P = 0.057 P < 0.008 P = 0.22 (D) ns P = 0.158 (D) ns ns

D=bilateral disadvantage; Nm=normal; SB=split brain; Ac=complete agenesis; Ap=partial agenesis.

lowed by a bilaterally similar positive wave peaking between 300 and 400 ms. As reported in Brown and Jeeves [11], for these 20 normal subjects there was a signi®cant visual ®eld-by-hemisphere interaction of peak latency for both P1 (F = 48.17, df=1/19, P < 0.001) and N1 (F = 81.27, df=1/19, P < 0.001). Cross-callosal di€erences in amplitude were also signi®cant (P1: F = 4.6, df=1/19, P < 0.05; N1: F = 36.06, df=1/19, P < 0.001). EPs averaged across two di€erent experimental runs for subject LB can also be seen in Fig. 1. P1 and N1 are clearly present for responses to stimulation in the contralateral visual ®eld (dashed line for left parietal, solid line for right parietal), although the N1 is considerably broader than the N1 seen in normal controls. However, for ipsilateral visual ®eld stimulation (solid line for left parietal, dashed line for right parietal), there is neither a P1 nor an N1 component evident in the response from either hemisphere. EP traces for two of the callosal agenesis subjects (MM and JD) can also be seen in Fig. 1. As with the commissurotomy subject (LB), both agenesis subjects have clear P1 and N1 components to contralateral visual ®eld stimulation, but neither component can be seen in the responses to ipsilateral stimulation. Similar results were seen in subject BE (not shown). What appears to be a very long latency negative wave is present in the LVF response recorded over the left hemisphere of JD. This negative wave does not appear to be an N1, but rather the end of a slow negative artifact beginning prior to stimulus onset. Absence of the

LVF-RH P1/N1 complex in JD was con®rmed by similar recordings during a simple RT task. A positive wave appears in the ipsilateral responses of MM, and to a lesser degree JD (left parietal response to a LVF stimulus) which is of somewhat longer peak latency (approximately 180±250 ms) than the contralateral N1 (150±180 ms). In addition, a later positive component can be seen in both contralateral and ipsilateral waveforms, but with longer latency for ipsilateral (300±500 ms). This component (perhaps a P3) does not appear in the waveforms of LB, but is consistently present in the responses of normal controls. Fig. 2 presents a similar pattern of EP waveforms for patient KF (partial callosal agenesis), including EPs from the three additional recording loci forming a coronal row. Again it can be seen that a large, clear P1/N1 complex occurs over the hemisphere contralateral to the visual ®eld of stimulation, but neither component can be seen in waveforms recorded over the hemisphere ipsilateral to the stimulus (i.e., no crosshemisphere P1 or N1). Partial agenesis patient AC had a similar pattern of EPs (not shown). As with MM and JD, a positive component appears in KF over the ipsilateral hemisphere with slightly longer latency than the contralateral hemisphere N1. Finally, the later positive wave again appears in KF with ipsilateral latency longer than contralateral. Fig. 2 includes a line graph of N1 amplitude for the coronal line of ®ve electrodes (measured against the average reference) which makes clear the hemispherically reciprocal nature of

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

the N1 response in the absence of the posterior 2/3 of the corpus callosum. 3.2. Letter and pattern matching performance 3.2.1. Adult normals Letter-matching performance for 20 normal young adults on this task is reported in Brown and Jeeves [11]. Performance for these normals (Table 2A, Adults) was well above chance level for both unilateral and bilateral presentations, with a signi®cant BFA for both error rate (F = 51.21, df=1/19, P < 0.001) and reaction time (F = 31,78, df=1/18, P < 0.001). These adult normal subjects were not run on the patternmatching task. 3.2.2. Adolescent boys Letter-matching performance for this group of subjects was similar to that of the adult normals (Table 2A, Boys). Performance was well above chance level for both unilateral and bilateral presentations, with a signi®cant BFA for both error rate (F = 9.74, df=1/ 14, P < 0.008) and reaction time (F = 13.56, df=1/14, P < 0.003). Pattern-matching performance in the adolescent boys also revealed a signi®cant bilateral advantage in error rate (F = 8.44, df=1/14, P < 0.012) and a strong trend for a bilateral advantage in RT (F = 4.29, df=1/14, P = 0.057; see Table 2B). 3.2.3. Commissurotomy patient LB Mean error rate and RTs for letter matching for commissurotomy patient LB can be seen in Table 2A. LB was not able to perform at a level exceeding chance on bilateral trials, but could do unilateral ®eld comparisons above chance level, that is, a signi®cant bilateral ®eld disadvantage (F = 29.74, df=1/1190, P < 0.001). Despite above-chance performance on unilateral trials, LB's performance on these trials was not as accurate as seen in normals in the same task. LB's reaction times also revealed a bilateral ®eld disadvantage (F = 4.37, df=1/1190, P < 0.037; Table 1A). For bilateral presentations a clear response bias was evident for bilateral horizontal stimuli which was not present for bilateral diagonal. That is, LB responded `match' to most bilateral horizontal presentations, yielding 12.7% error for match trials and 85.1% error for no-match trials. For bilateral diagonal presentations, error rates were around 50% for both diagonal trial types. A position (horizontal vs diagonal) by stimulus (match vs no-match) ANOVA of bilateral response accuracy yielded signi®cant stimulus (F = 43.31, df=1/347, P < 0.001) and position-bystimulus e€ects (F = 40.61, df=1/347, P < 0.001). LB was not tested on the pattern-matching task. Based on his letter-matching performance, it was pre-

1173

sumed that he would not have been able to make bilateral pattern matches above the level of chance (i.e., a substantial bilateral disadvantage). 3.2.4. Complete callosal agenesis Table 2A presents the letter-matching performance for the four complete agenesis patients (JD, MM, BE and MO). All performed well above chance levels for bilateral letter matching. Two of the four had some evidence of a bilateral advantage (in the case of BE the bilateral advantage was signi®cant). JD's overall performance was so accurate that a bilateral advantage would have been dicult to demonstrate. Notably, despite callosal absence JD was more accurate on bilateral trials than the average normal (Adults and Boys). Reaction times for letter matching indicated that the accurate performance of the complete ACC patients did not come at a price in response time. Only MO had an average RT outside the range of normals, but he was also slower on unilateral trials. Two of the four had faster average RTs for bilateral trials than unilateral (MM showing a signi®cant bilateral advantage). Pattern-matching data for these four complete ACC patients can be seen in Table 2B. Here the results appear to di€er from normal. All four of these complete ACC patients had error rates for bilateral trials that were outside the range for normals, whereas their unilateral performance was consistent with that of normals. Three of four complete ACC patients had substantial bilateral disadvantages (D) in error rate, two of these being signi®cant disadvantages. RTs for the pattern-matching task were little di€erent than those of normal, with the exception of MO's overall slow response time. There was no evidence of bilateral advantages in RT among these four individuals, with MM showing a trend toward a bilateral disadvantage. In summary, complete ACC patients can clearly make bilateral comparisons of visual letters at a level not di€erent from normals. However, they seem to have relatively greater diculty with bilateral matching of pattern stimuli compared to normals, although performance exceeded chance. 3.2.5. Partial callosal agenesis The results of unilateral and bilateral matching for individuals with partial callosal agenesis (KF and AC) can also be seen in Table 2. For letter matching, both of these individuals were more accurate than the average for normals (Adults and Boys). AC was so accurate on unilateral trials there was little room for a bilateral advantage, while KF showed a statistically signi®cant bilateral advantage in errors. This lettermatching accuracy for bilateral trials was accomplished at little sacri®ce in speed, as evidenced by bilateral RTs not markedly slower than normal individuals.

1174

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

Partial ACC patients also performed normally on pattern matching. While the bilateral error rate increased for both relative to the letter task, they were both more accurate than the average for normal adolescents. Despite normal accuracy, AC did show a signi®cant bilateral disadvantage in error rate. RTs for both partial ACC individuals were longer than normal. As with letter-matching, KF had a bilateral advantage in RT, while AC showed a weak trend for a bilateral disadvantage in RT for pattern matching. In summary, individuals with partial agenesis appeared to do at least as well as the normal groups in the letter task. In the pattern task, the loss of bilateral matching accuracy (relative to letter matching) which was seen in the complete ACC was less severe in the data of partial ACC patients. The two partial ACC individuals were overall more accurate at matching tasks than the normal subjects.

4. Discussion These results support the following conclusions: (1) Both acallosals (partial and complete) and the commissurotomy patient show no evidence of the presence of either P1 or N1 components in EPs recorded over the hemisphere ipsilateral to visual stimulation. (2) Unlike the commissurotomy patient, acallosals appear to have (in most cases) a positive wave in ipsilateral waveforms that is nearly concomitant with the contralateral N1, but with slightly longer latency. Acallosals, but not the commissurotomy patient, also have a later positive wave in the EP that is recordable over both the ipsilateral and contralateral hemispheres, again with longer latency in ipsilateral records. (3) While the commissurotomy patient performed at chance level for crossvisual ®eld letter matching (a bilateral disadvantage), all six individuals with callosal agenesis (partial and complete) were able to perform above chance level for bilateral ®eld letter comparisons, with a BFA occurring in 3 of 6 patients. (4) For the pattern-matching task, complete agenesis patients showed a drop in overall performance and had much greater diculty with bilateral than unilateral matches (bilateral disadvantages), while the partial agenesis patients maintained more accurate bilateral pattern-matching performance (although partial acallosals were also more accurate overall). (5) The commissurotomy patient showed a strong bias for responding `match' to horizontal-bilateral presentations (letters in homologous locations in the two visual ®elds), but no such bias to bilateral diagonal presentations. This pattern of response bias was not seen in the responses of the acallosals

4.1. Cross-callosal EPs Absence of the cross-callosal P1/N1 components of the EP in the patients tested is a replication of the results reported by Rugg et al. [62] for complete agenesis patients, and by Mangun et al. [43] for the commissurotomy patient. The EP data reported herein are unique in the EP literature on callosal function in allowing a direct comparison of EPs from acallosals, commissurotomy patients, and normals using the same recording methods, task, and parameters of stimulus presentation. In a reanalysis of the EPs from MM, JD and LB (as well as another commissurotomy patient, NG), Brown et al. [10] demonstrated that even with the greater signal extraction power of latency-adjusted EP averaging, ipsilateral (cross-callosal) P1/N1 components could not be detected. Speiser et al. [72] report dense array ®eld mapping of visual evoked potentials to lateralized stimuli in subjects MM and JD reported here. Even with dense array topographic ®eld mapping, neither P1 nor N1 were evident over the hemisphere ipsilateral to ®eld of visual stimulation. Most importantly, previous studies had not presented EP data in individuals with partial ACC. The pattern of EP results was remarkably similar in our partial agenesis patients (lacking the middle and posterior callosum) to that seen in the complete agenesis or commissurotomy patients. Similar results are reported by Marzi et al. [46] for EP recordings in two patients with traumatic lesions of the corpus callosum. A patient with lesion of the posterior third of the corpus callosum showed no evidence of an ipsilateral P1/N1 complex to unilateral visual stimulation. The other patient, whose CT revealed a hemorragic lesion in the middle third of the corpus callosum, showed clear evidence of a bilateral P1/N1 complex in the visual EPs. Tramo et al. [78] also report a patient in whom transection of the anterior 2/ 3 of the corpus callosum did not eliminate the ipsilateral P1/N1 complex to a unilateral visual stimulus. However, a later transection of the posterior body and splenium resulted in elimination of the ipsilateral P1/ N1. Absence of a cross-callosal P1/N1 in both the commissurotomy and acallosal patients constitutes strong evidence that these same components, when recorded in normals over the hemisphere ipsilateral to the visual ®eld of stimulation, are a product of callosal transmission and regeneration within ipsilateral cortex (see also [62]). Thus, the di€erence in latency between contralateral and ipsilateral recordings can correctly be considered a reasonable estimate of the time required for evoked potential interhemispheric transmission in normals [11,12,35,59,60,65], as well as a measure of posterior callosal integrity [13,17,44]. The absence of an ipsilateral P1/N1 in the partial agenesis patient

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

speci®cally implicates the posterior callosum (likely the splenium) as the location of transfer of the neural activity that results in an ipsilateral P1/N1 complex. These EP data are also relevant to question of the possible role of the anterior commissure in interhemispheric interactions. The anterior commissure, found to be clearly present in all four complete agenesis patients, was not sucient to mediate the cross-callosal visual P1/N1 complex. Nevertheless, the appearance of a slightly later positive wave in recordings over the ipsilateral hemispheres in acallosals suggests that the anterior commissure is mediating sucient interhemispheric interaction to elicit some electrical response in ipsilateral cortex. There is what appears to be a similar positive wave in the left parietal response to a LVF stimulus in LB. However, in the case of LB the positive wave does not have the same delayed latency as seen in the ACC individuals, a similar wave does not appear in the right parietal response to a RVF stimulus, and the positive wave does not appear consistently in subaverages from the various trial blocks. Thus, it appears that this positive wave in LB is most likely an artifact, and therefore di€erent than that seen in the cases of complete or partial ACC. Saron et al. [66] presented results that support the hypothesis that some form of ipsilateral visual EP activity may be mediated by the anterior commissure. They recorded dense array visual evoked potential maps of RVF and LVF response in one ACC individual with an anterior commissure and one without an anterior commissure. While the electrical map foci indicating the P1/N1 complex could not be seen ipsilaterally in either case, only in the individual with an anterior commissure could a somewhat later (180±200 ms) occipital focus be seen over the hemisphere ipsilateral to the ®eld of stimulation. Saron et al. [66] conclude that this ipsilateral focus represents activity resulting from interhemispheric in¯uences arriving via the anterior commissure. This component in acallosals may be a P2 component occurring over ipsilateral cortex in the absence of an overlying P1/N1 complex. Another di€erence between the ipsilateral waveforms of the commissurotomy patient and the acallosals was the appearance of a late positive wave (>300 ms) in the responses of normals, and acallosals, but not the commissurotomy patient. Again, this wave was of longer latency over ipsilateral than contralateral cortex. This component appears to be a P3. Occurrence of a P3 over ipsilateral cortex would be consistent with the fact that information regarding the visual stimulus does reach the other hemisphere in acallosals (allowing for above chance matching in both the letter and pattern tasks), but not in the commissurotomy patient (suggested by the inability to do the bilateral letter matching task). Thus, EPs suggest that the posterior

1175

callosum is necessary for transfer of the sensory information represented by the P1/N1 complex. However, other cerebral commissures (likely the anterior commissure) present in ACC individuals allow sucient information transfer for elicitation of these two later positive waves. Innocenti [29] suggests that in partial callosal agenesis the anterior callosum (when present) may carry ®bers that would normally cross in more posterior portions of the callosum. Similarly, Auroux and Roussel [1] demonstrated that the anterior commissure interconnects widespread neocortical areas around the 10th gestational week, but by the 16th week most of these exuberant connections have contracted as they are replaced by callosal ®bers. It is possible that callosal absence in ACC might allow for greater retention of the exuberant connections through the anterior commissure. Our data demonstrate that, if ®bers interconnecting posterior visual areas are present in the anterior callosal remnant of partial ACC or anterior commissure of complete ACC, they are not sucient to elicit the P1/N1 component complex in ipsilateral posterior visual cortex, although they might account for the later positive components which occur over ipsilateral cortex. 4.2. Bilateral visual ®eld processing 4.2.1. Letter matching Despite a similar absence of the P1/N1 over the ipsilateral hemisphere in both the commissurotomy and the acallosal patients (complete and partial), there are distinct di€erences between these patients in bilateral letter-matching performance. LB was unable to perform above chance on bilateral letter matching, while all ACC individuals were clearly above chance on these trial types. For the letter task, all acallosals except MO showed a signi®cant BFA in either accuracy or reaction time. It has been demonstrated that cross-uncrossed RT di€erences (CUDs) [4] are prolonged in ACC, and even more so in commissurotomy patients [6]. However, crossed responses are never eliminated, even in complete commissurotomy patients. In addition, Reuter-Lorenz et al. [56] demonstrate that the redundant target e€ect (i.e., faster simple RT to bilateral than to unilateral presentations) occurs even in complete commissurotomy patients. These studies suggest that subcortical systems are sucient for a crossed response to stimulus occurrence and for crossed enhancement of RTs in this task. Marzi et al. [46] suggest that there are at least two channels available for mediating the CUD, a faster callosal channel and a slower subcortical response channel, and that the ipsilateral P1/N1 complex may be an index of the operation of the faster callosal channel.

1176

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

Both the current data from letter and pattern matching, and data from the CUD task reported by both Saron et al. [66] and Marzi et al. [46] demonstrate a dissociation between EPs (at least with respect to the P1/N1 complex) and behavioral performance. However, the CUD task requires callosal transmission of the occurrence of a visual stimulus only, not the interhemispheric transfer of the content of the stimulus required by the bilateral letter and pattern matching tasks. Therefore, although the results of these two tasks in commissurotomy patients and acallosals can be compared to some degree, the two tasks represent substantially di€erent challenges for interhemispheric transfer. The inability of the commissurotomy patient to do bilateral ®eld matching above a chance level when comparisons involve more than a simple visual feature has been a consistent aspect of the disconnection syndrome in these patients [16,34,76±78]; although Johnson [34] and Eviatar and Zaidel [21] present a possible exception in one split-brain patient. The fact that the complete agenesis patients were clearly above chance in bilateral letter matching strongly suggests the possibility that information is being compared via the anterior commissure which was visible in the MRIs of all four. An alternative possibility is that interhemispheric transmission occurs via the posterior commissure, but this structure could not be reliably visualized in the MRIs of these patients. Finally, the hypothesis of subcortical transfer might be ruled out since bilateral letter comparisons should also have been possible on this basis in the split-brain patient. Similarly, MRI ®ndings in LB revealed the presence of a massa intermedia [8] suggesting that this structure was also not able to mediate letter comparisons in LB. In the case of callosal agenesis, it is possible (though perhaps unlikely) that developmental compensation for callosal absence may create conditions that allow for subcortical transfer in agenesis that are not present in the commissurotomy patient. The role of the anterior commissure as a compensatory mechanism allowing for interhemispheric transfer in acallosals is not yet clearly understood. In the macaque, Hamilton and Vermeire [28] showed that interhemispheric transfer of visual discriminations based on pattern, orientation, movement and facial feature can take place via either the splenium of the corpus callosum or the anterior commissure. Risse, et al [58] and McKeever and Sullivan [47] report con¯icting results with respect to the capacity of the anterior commissure to mediate interhemispheric transfer in human patients with callosotomies that spared the anterior commissure. Both papers report a failure of tactile transfer in these patients. Risse et al. report retained visual transfer in 3 of 4 patients, while McKeever and Sullivan report no evidence of visual transfer in 2 patients

tested using similar methods. Thus, while tactile transfer was consistently absent, visual transfer may or may not occur across the anterior commissure. Particulars of other neuropathology, disease history, surgery, and/ or subject strategy likely account for the di€erences in the success of visual transfer in these studies. Finally, Tramo et al. [78] report an inability to match the orientation, form or luminance of bilaterally presented visual stimuli in patients with transections of the corpus callosum and posterior commissure, but sparing the anterior commissure. Chiarello [14], in a review of cognitive processing in 29 reported cases of ACC, emphasized an important role of the anterior commissure, at least in the transfer of less complex sensory information. Fischer et al. [23] speci®cally studied the role of the anterior commissure in one acallosal in whom the anterior commissure could not be convincingly identi®ed in the MRI, compared to an ACC individual with a hypertrophied anterior commissure. These investigators conclude, on the basis of lesser de®cits in interhemispheric transfer in the case of hypertrophy, that the anterior commissure is an important mechanism in interhemispheric transfer in ACC. However, the mild de®cits in the patient with apparently little or no anterior commissure suggested that other compensatory mechanisms or interhemispheric pathways might also be involved. 4.2.2. Pattern matching A comparison between the results of bilateral lettermatching compared to bilateral pattern-matching in the acallosal patients is particularly revealing with regard to the role of the corpus callosum in bilateral visual ®eld integration. Individuals with complete agenesis performed relatively normally, i.e. showed evidence of a BFA, on the letter-matching task, but showed signi®cant de®cits in bilateral visual comparisons in the case of patterns (i.e., a bilateral disadvantage). This di€erence suggests that letter information could transfer reliably but the pattern information could not. A similar result was reported by Ettlinger et al. [20] who found that individuals with total agenesis showed de®cits in the ability to match bilaterally presented dot patterns, but were relatively normal at bilateral matching of shades of green or grey. Lassonde [36] has reported results of same/di€erent judgements of letters, numbers, colors, and forms in 6 ACC patients. With regard to accuracy, Lassonde concludes that the ACC patients were able to do interhemispheric matching as well as they could do intrahemispheric matching, and as well as controls. With respect to response latency, acallosals took twice as much time to respond as the control group. Although not analysed by Lassonde, the means showed a slight bilateral advantage in RT for normals, and a slight bilateral disadvantage for the acallosals.

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

However, Lassonde did not report separately data for letters and forms, so it is dicult to compare her data directly to those reported herein. Di€erences in the outcomes of these two experiments may be related to stimulus exposure duration, which was faster (60 ms) in this experiment than that used by Lassonde (150 ms). This suggests that alternative, noncallosal pathways may require a greater amount of time for sucient information to cross in order to allow for an accurate matching response. Finally, the potential for di€erent choices regarding a speed/accuracy trade-o€ means that detecting the disability of an ACC individual requires analysis of both accuracy and reaction time. The letter and pattern stimuli used in this study differed in the amount of spatial information, in complexity, and in verbal encodability, with pattern stimuli being more spatial, more complex, and more dicult to assign a verbal label. Each of these properties has received some attention with respect to access to interhemispheric transfer via the anterior commissure. Martin [45] argued that the anterior commissure might be capable of transferring pattern but not spatial information. However, the lack of transfer of spatial information in acallosals with an anterior commissure could not be replicated by Jeeves and Milner [33]. Chairello [14], in her review of acallosal cases, emphasized complexity, suggesting that `the ability to cross integrate complex visual information is beyond the capacities of the acallosal brain' (p. 143). The importance of stimulus complexity in the interhemispheric transfer of information in the agenesis patient is illustrated in the results of interhand transfer of the tactile-spatial information of the Form Board (Tactile Performance Test) described earlier. Sauerwein et al. [69] reported unimpaired performance in two asymptomatic individuals with agenesis of the corpus callosum using a 6-block TPT. Fischer et al. [23] found only mild impairment using the intermediate (8-block) version. Finally, Sauerwein et al. [68] reported in their more recent work impaired performance on the 10block TPT in the same two acallosals. Paul and Brown [53] also report de®cits in acallosals on the 10-block TPT that closely mirror the more recent data of Sauerwein et al. It can be concluded from these TPT data that diculty (number of blocks) is an important variable in the appearance of a de®cit in interhemispheric transfer in the individual with ACC. That is, while information relative to a simpli®ed version of the TPT (or the letters of our task) may cross via intact noncallosal channels, information regarding the more complex 10-block TPT (or complex visual patterns) cannot cross with sucient integrity to sustain normal interhemispheric performance. The role of complexity in the interhemispheric transfer of tactile information in ACC patients is also clear in the work of Ge€en et

1177

al. [25]. These investigators report greater de®cits in inter-hand transfer of ®nger localization in acallosals for two- and three-®nger sequences than are seen for single ®nger stimulation. Gazzaniga et al. [24] speculate that encodability is an important variable in the interhemispheric transfer of stimulus information. These authors suggest that `the initial decoding of primary sensory input, such as that which must occur in response to linguistic stimuli, renders the information more `transferable' than its original form.' The dot patterns used in this study may have produced greater transfer de®cits in the acallosals because they were dicult to encode. 4.2.3. Partial agenesis Patient KF did not show the relatively greater diculty making bilateral comparisons of patterns than letters that was so marked in the results of complete agenesis patients. In contrast, partial agenesis patient AC was more a€ected by patterns, increasing in bilateral errors and manifesting a bilateral disadvantage. However, AC's error rate for bilateral pattern matching was slightly below that of the normal group, and well below that of the complete agenesis patients. Thus, even AC seemed to be able to eciently match bilaterally presented patterns. Ettlinger et al. [20] also found that individuals with partial agenesis did not show de®cits in bilateral dot-pattern matching. These ®ndings in partial agenesis suggest that the presence of the genu of the corpus callosum is sucient to sustain near-normal interhemispheric comparisons of more complex visual stimuli. Thus, the contribution of residual callosal ®bers to interhemispheric interactions becomes increasingly apparent as complexity of the stimulus information to be transferred increases. Either anterior or posterior portions of the callosum seem to be sucient for adequate interhemispheric transfer. Gordon et al. [26] reported the results of tests of interhemispheric transfer in 2 patients with transections of the anterior commissure and anterior 2/3 of the corpus callosum. No de®cits could be found in interhemispheric transfer of visual, tactile, motor or complex spatial (formboard) learning. Thus, a small portion of posterior callosum can mediate interhemispheric sensory integration without the participation of the anterior commissure or anterior callosum. It is uncertain whether the genu of the corpus callosum remaining in the partial agenesis individuals contains the same ®bers as those which would be present in a normally developing callosum, or, as Innocenti [29] has argued, also contains ®bers interconnecting cortical areas normally interacting through more posterior callosal paths. However, for the letter and pattern stimuli of this study, the existing genu was sucient to sustain higher levels of bilateral lettermatching performance.

1178

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

4.2.4. Homologous visual ®ll-in LB showed a strong response bias for responding `match' for bilateral horizontal trials that was not seen for bilateral diagonal trials. This bias may be related to the ®ll-in phenomenon demonstrated by commissurotomy patients in the chimeric faces experiment [39]. That is, whenever there is a stimulus of some kind occupying the homologous location in the opposite visual ®eld, each hemisphere of the commissurotomy patient presumes the information in the other hemisphere to be identical to that available in its own ®eld of view. Thus, every horizontally presented letter pair appeared to LB to be a match. Interestingly, at the end of the experimental session LB volunteered that he felt he had got all of the bilateral horizontal comparisons correct, but felt he had had trouble with the bilateral diagonals. In fact he had performed at chance level on both bilateral horizontals and diagonals. Thus, the perceptual ®ll-in was subjectively convincing to LB. Information regarding the occurrence of a visual stimulus at a homologous position in space may be available to both hemispheres via subcortical (e.g., tectal) commissures in split-brain patients, but without further information the stimulus known to each hemisphere is perceptually ®lled into the ipsilateral ®eld percept held by that hemisphere. In the acallosal, the still-present anterior commissure must play a role in reducing this ®ll-in by providing sucient information regarding the nature of the actual stimulus so that the acallosals can make above chance bilateral visual ®eld comparisons and inhibit any bias toward `match' responses on bilateral horizontal trials.

Acknowledgements This research was supported in part by NICHD grant No. 1 R15 HD33118-01A1. The authors would like to thank Dr Joseph Bogen for his very helpful comments on this manuscript.

References [1] Auroux M, Roussel C. Relations neocorticales assureÁes par la commisure blanche anteÁrieure chez le foetus human. CR Assoc Anat 1967;137:152±9. [2] Baierl P, Forster C, Fendel H, Haegele M, Fink U, Kenn W. Magnetic resonance imaging of normal and pathological white matter maturation. Pediatric Radiology 1988;18:183±9. [3] Banich MT, Belger A. Interhemispheric interaction: How do the hemispheres divide and conquer a task? Cortex 1990;26:77±94. [4] Bashore TR. Vocal and manual reaction time estimates of interhemispheric transmission time. Psychological Bulletin 1981; 89:352±68. [5] Berger JM. Interhemispheric cooperation and activation in integration of verbal information. Behavioural Brain Research 1988;29:193±200.

[6] Berlucchi G, Aglioti S, Marzi CA, Tassinarie G. Corpus callosum and simple visuomotor integration. Neuropsychologia 1995;33:923±36. [7] Bogen JE. The other side of the brain. I: Dysgraphia and dyscopia following cerebral commissurotomy. Bulletin of the Los Angeles Neurological Society 1969;34:73±105. [8] Bogen JE, Schultz DH, Vogel PJ. Completeness of callosotomy shown by magnetic resonance imaging in the long term. Archives of Neurology 1988;45:1203±5. [9] Bogen JE, Vogel PJ. Cerebral commissurotomy in man. Bulletin of the Los Angeles Neurological Society 1962;27:169±72. [10] Brown WS, Bjerke MD, Galbraith GC. Interhemispheric transfer in normals and acallosals: Latency adjusted evoked potential averaging, Cortex 1998;34:677±692. [11] Brown WS, Jeeves MA. Bilateral ®eld advantage and evoked potential interhemispheric transmission time. Neuropsychologia 1993;31(12):1267±81. [12] Brown WS, Larson EB, Jeeves MA. Directional asymmetries in interhemispheric transfer time: Evidence from visual evoked potentials. Neuropsychologia 1994;32:439±48. [13] Burnison DS, Larson EB, Brown WS. Evoked potential interhemispheric transmission and the bilateral ®eld advantage in Multiple Sclerosis. Journal of the International Neuropsychological Society 1995;1:182 abstract. [14] Chiarello C. A house divided? Cognitive functioning in callosal agenesis. Brain and Language 1980;11:128±58. [15] Corballis MC. Visual integration in the split brain. Neuropsychologia 1995;33:937±59. [16] Corballis MC, Trudel CI. Role of the forebrain commissures in interhemispheric integration. Neuropsychology 1993;7:306±24. [17] Davidson RJ, Saron CD. Evoked potential measures of interhemispheric transfer time in reading disabled and normal boys. Developmental Neuropsychology 1992;8:261±77. [18] Dimond S, Beaumont G. Use of two cerebral hemispheres to increase brain capacity. Nature 1971;232:270±1. [19] Dimond S, Beaumont G. Processing in perceptual integration between and within the cerebral hemispheres. British Journal of Psychology 1972;63:509±14. [20] Ettlinger G, Blakemore CB, Milner AD, Wilson J. Agenesis of the corpus callosum: A behavioral investigation. Brain 1972; 95:327±46. [21] Eviatar Z, Zaidel E. Letter matching within and between the disconnected hemispheres. Brain and Cognition 1994;25:128±37. [22] Ferriss GS, Dorsen MM. Agenesis of the corpus callosum 1. Neuropsychological studies. Cortex 1975;11:95±122. [23] Fischer M, Ryan SB, Dobyns WB. Mechanisms of interhemispheric transfer and patterns of cognitive function in acallosal patients of normal intelligence. Archives of Neurology 1992;49:271±7. [24] Gazzaniga MS, Risse GL, Springer SP, Clark E, Wilson DH. Psychologic and neurologic consequences of partial and complete cerebral commissurotomy. Neurology 1975;25:10±5. [25] Ge€en GM, Nilsson J, Simpson DA, Jeeves MA. The development of interhemispheric transfer of tactile information in cases of callosal agenesis. In: Lassonde M, Jeeves MA, editors. Callosal agenesis: A natural split-brain? New York: Plenum Press, 1994. [26] Gordon WH, Bogen JE, Sperry RW. Absence of deconnexion syndrome in two patients with partial section of the neocommissures. Brain 1971;94:327±36. [27] Gott PS, Saul RE. Agenesis of the corpus callosum: Limitations of functional compensation. Neurology 1978;28:1272±9. [28] Hamilton CR, Vermeire BA. Localization of visual functions with partially split-brain monkeys. In: Lepore F, Ptito M, Jasper HH, editors. Two hemispheresÐone brain: functions of the corpus callosum. New York: Alan R. Liss, 1986. p. 315±33. [29] Innocenti GM, Aggoun-Zouaoui D, Lehmann P. Cellular

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

[30]

[31]

[32] [33]

[34] [35] [36]

[37] [38]

[39] [40] [41] [42] [43]

[44]

[45] [46]

[47]

[48]

aspects of callosal connections and their development. Neuropsychologia 1995;33:961±87. Jeeves MA. Some limits to interhemispheric integration in cases of callosal agenesis and partial commissurotomy. In: Russell I, van Hof M, Berlucchi G, editors. Structure and function of the cerebral commissures. New York: Macmillan, 1979. p. 449±74. Jeeves MA. Callosal agenesis: Neuronal and developmental adaptations. In: Lepore F, Ptito M, Jasper HH, editors. Two hemispheresÐone brain: functions of the corpus callosum. New York: Alan R. Liss, 1986. p. 403±21. Jeeves MA, Lamb A. Cerebral asymmetries and interhemispheric processes. Behavioural Brain Research 1988;29:211±23. Jeeves MA, Milner AD. Speci®city and plasticity in interhemispheric integration: Evidence from callosal agenesis. In: Ottoson D, editor. The duality and unity of the brain. London: Macmillan, 1987. Johnson LE. Bilateral visual cross-integration by human forebrain commissurotomy subjects. Neuropsychologia 1984;22 (2):167±75. Larson EB, Brown WS. Bilateral ®eld interactions, hemispheric specialization and evoked potential interhemispheric transmission time. Neuropsychologia 1997;35:573±81. Lassonde M. The facilitatory in¯uence of the corpus callosum on interhemispheric processing. In: Lepore F, Ptito M, Jasper HH, editors. Two hemispheresÐone brain: functions of the corpus callosum. New York: Alan R. Liss, 1986. Lassonde M, Jeeves MA, editors. Callosal agenesis: A natural split-brain? New York: Plenum Press, 1994. Lassonde M, Sauerwein H, McCabe N, Laurencelle L, Geo€ry G. Extent and limits of cerebral adjustment to early section or congenital absence of the corpus callosum. Behavioral Brain Research 1988;30:165±81. Levy J, Trevarthan C, Sperry RW. Perception of bilateral chimeric ®gures following hemispheric deconnexion. Brain 1972;95:61±78. Liederman J, Merola J, Martinez S. Inter-hemispheric collaboration in response to simultaneous bilateral input. Neurosychologia 1985;23:673±83. Lines CR, Rugg MD, Milner AD. The e€ect of stimulus intensity on visual evoked potential estimates of interhemispheric transmission time. Experimental Brain Research 1984;57:89±98. Ludwig TE, Jeeves MA, Norman WD, DeWitt R. The bilateral ®eld advantage on a letter-matching task. Cortex 1993;29:691± 713. Mangun GR, Luck SJ, Gazzaniga MS, Hillyard SA. Electrophysiological measures of interhemispheric transfer of visual information: Studies of split-brain patients. Society for Neuroscience 1991;17:340 (4). Markee TE, Brown WS, Moore LH, Theberge DC, Zvi JC. Callosal function in dyslexia: Evoked potential interhemispheric transfer time and bilateral ®eld advantage. Developmental Neuropsychology 1996;12:409±28. Martin A. A qualitative limitation on visual transfer via the anterior commissure. Brain 1985;108:43±63. Marzi CA, Bongiovanni M, Miniussi C, Samnia N. E€ects of partial callosal and unilateral cortical lesions on interhemispheric transfer. In: Zaidel E, Iacoboni M, Pascual-Leone A, editors. The corpus callosum in sensory motor integration: individual di€erences and clinical applications, In press. McKeever WF, Sullivan KF. Typical cerebral hemisphere disconnection de®cits following corpus callosum section despite sparing of the anterior commissure. Neuropsychologia 1981;19:745±55. Merola J, Liederman J. The extent to which between-hemisphere division of inputs improves performance depends upon task diculty. Journal of Clinical and Experimental Neuropsychology 1990;10:69.

1179

[49] Milner AD. Visual integration in callosal agenesis. In: Lassonde M, Jeeves MA, editors. Callosal agenesis: A natural split-brain? New York: Plenum Press, 1994. [50] Myers JJ, Sperry RW. Interhemispheric communication after section of the forebrain commissures. Cortex 1985;21:249±60. [51] Norman WD, Jeeves MA, Milne A, Ludwig T. Hemispheric interactions: The bilateral advantage and task diculty. Cortex 1992;28:623±42. [52] Old®eld RC. The assessment of handedness: The Edinburgh inventory. Neuropsychologia 1971;9:97±113. [53] Paul AL, Brown WS. Psychosocial de®cits in agenesis of the corpus callosum. Journal of the International Neuropsychological Society 1996;2:41 abstract. [54] Posner MI, Mitchell RF. Chronometric analysis of classi®cation. Psychological Review 1967;74:392±409. [55] Rauch RA, Jinkins JR. Magnetic resonance imaging of corpus callosum dysgenesis. In: Lassonde M, Jeeves MA, editors. Callosal agenesis: A natural split-brain? New York: Plenum Press, 1994. p. 83±95. [56] Reuter-Lorenz PA, Nozawa G, Gazzaniga MS, Hughes HC. Fate of neglected targets: A chronometric analysis of redundant target e€ects in the bisected brain. Journal of Experimental Psychology: HPP 1995;21:211±30. [57] Reynolds DM, Jeeves MA. Further studies of tactile perception and motor coordination in agenesis of the corpus callosum. Cortex 1977;13:257±72. [58] Risse GL, LeDoux J, Springer SP, Wilson DH, Gazzaniga MS. The anterior commissure in man: Functional variation in a multisensory system. Neuropsychologia 1978;16:23±31. [59] Rugg MD, Beaumont G. Interhemispheric asymmetries in the visual evoked response: E€ects of stimulus lateralisation and task. Biological Psychology 1978;6:283±92. [60] Rugg MD, Lines CR, Milner AD. Visual evoked potentials to lateralized visual stimuli and the measurement of interhemispheric transmission time. Neuropsychologia 1984;22:215± 25. [61] Rugg MD, Lines CR, Milner AD. Further investigation of visual evoked potentials elicited by lateralized stimuli: E€ects of stimulus eccentricity and reference site. Electroencephalography and Clinical Neurophysiology 1985;62:81±7. [62] Rugg MD, Milner AD, Lines CR. Visual evoked potentials to lateralised stimuli in two cases of callosal agenesis. Journal of Neurology, Neurosurgery, and Psychiatry 1985;48:367±73. [63] Russell JR, Reitan RM. Psychological abnormalities in agenesis of the corpus callosum. Journal of Nervous and Mental Disease 1955;121:205±14. [64] Salamy A. Commissural transmission: Maturational changes in humans. Science 1978;200:1409±10. [65] Saron CD, Davidson RJ. Visual evoked potential measures of interhemispheric transfer time in humans. Behavioral Neuroscience 1989;103:1115±38. [66] Saron CD, Lassonde M, Vaughan HG, Foxe JJ, Alfhors SP, Simpson GV. Interhemispheric visuomotor interaction in callosal agenesis: Spatiotemporal patterns of cortical activation. Society for Neuroscience Abstracts 1997;23:1949 abstract. [67] Sauerwein H, Lassonde MC. Intra- and interhemispheric processing of visual information in callosal agenesis. Neuropsychologia 1983;21(2):167±71. [68] Sauerwein HC, Lassonde M. Cognitive and sensori-motor functioning in the absence of the corpus callosum: neuropsychological studies in callosal agenesis and callosotomized patients. Behavioural Brain Research 1994;64:229±40. [69] Sauerwein HC, Lassonde M, Cardu B, Geo€roy G. Interhemispheric integration of sensory and motor functions in callosal agenesis. Neuropsychologia 1981;19:445±54. [70] Sergent J. A new look at the human split-brain. Brain 1987; 110:1375±92.

1180

W.S. Brown et al. / Neuropsychologia 37 (1999) 1165±1180

[71] Solursh LP, Margulies AI, Ashem B, Stasiak EA. The relationship of agenesis of the corpus callosum to perception and learning. Journal of Nervous and Mental Disease 1965;141:180± 9. [72] Speiser AH, Tucker DM, McDougal L, Murias M, Brown WS. Absent electrical responses in the unstimulated hemisphere in agenesis of the corpus callosum. In: Zaidel E, Iacoboni M, editors. The role of the human corpus callosum in sensory motor integration: NATO Advanced Study Institute (in press). [73] Sperry RW. Hemisphere deconnection and unity in conscious awareness. American Psychologist 1968;23:723±33. [74] Sperry RW. Some e€ects of disconnecting the cerebral hemispheres. Science 1982;217:1223±6. [75] Sperry RW. Consciousness, personal identity, and the divided brain. In: Benson DF, Zaidel E, editors. The dual brain: Hemispheric specialization in humans. New York: Guilford Press, 1985. [76] Sperry RW, Gazzaniga MS, Bogen JE. Interhemispheric relationships: The neocortical commissures; syndromes of hemisphere disconnection. In: Vinken PJ, Bruyn GW, editors.

[77]

[78]

[79]

[80]

Handbook of Clinical Neurology, Vol. 4. Amsterdam: Elsevier, 1969. p. 273±90. Seymor SE, Reuter-Lorenz PA, Gazzaniza MS. The disconnection syndrome: Basic ®ndings rearmed. Brain 1994;117:105± 15. Tramo MJ, Baynes K, Fendrich R, Mangun GR, Phelps EA, Reuter-Lorenz PA, Gazzaniga MS. Hemispheric specialization and interhemispheric integration: Insights from experiments with commissurotomy patients. In: Reeves AG, Roberts DW, editors. Epilepsy and the corpus callosum II. New York: Plenum Press, 1995. p. 263±95. Thompson KM, Amano SA, Khatchikian A, Brown WS. Development of the corpus callosum: Interhemispheric transmission time and bilateral ®eld advantage. Journal of the International Neuropsychological Society 1997;61:3 abstract. Yakovlev PI, Lecours AR. The myelogenetic cycles of regional maturation in the brain. In: Minowski A, editor. Regional development of the brain in early life. Oxford: Blackwell, 1967. p. 98±114.