Brief communication Contralateral visual search deficits following TMS

Transcription

1 501 Journal of Neuropsychology (2008), 2, q 2008 The British Psychological Society The British Psychological Society Brief communication Contralateral visual search deficits following TMS I. Schindler 1,A.Ellison 2 and A. D. Milner 2 * 1 Department of Psychology, University of Hull, UK 2 Cognitive Neuroscience Research Unit, Wolfson Research Institute, University of Durham, UK Transcranial magnetic stimulation of posterior parietal vs. superior temporal sites cause differential effects on conventional conjunction vs. feature search tasks, respectively. We now report that when adecision has to be made on the target s left/right location, different lateralized deficits emerge in these two cases. With full-field arrays, we found a specific PPC search deficit for contralateral space. With smaller, structured arrays presented in left or right hemispace, we found a specific STG deficit for contralateral parts of the array. Transcranial magnetic stimulation (TMS) of the right posterior parietal cortex (PPC) impairs difficult (serial) search for visual conjunctions, but not difficult feature search. In contrast, TMS of right superior temporal gyrus (STG) impairs difficult feature search, but not conjunction search(ellison, Schindler,Pattison, &Milner,2004). A similar deficit has been found to follow electrical stimulation of the STG (Gharabaghi, Fruhmann Berger, Tatagiba, & Karnath, 2006). These findings showed that disruption of either area can cause deficits in visual exploration, but that these deficits are task dependent. Our workhad been stimulated by the currentdebate as to which of these two regions forms the core substrate for visual neglect symptoms (Karnath, Ferber, & Himmelbach, 2001; Karnath, Fruhmann Berger, Kuker, & Rorden, 2004; Mort et al., 2003, Vallar & Perani, 1986), given that a fundamental aspect of neglect is a lateralized failure in exploratory search (Karnath, Milner, & Vallar, 2002). While not directly bearing on this debate, our data did provide independent evidence that both of these areas are involved, in different ways, in guiding visual search in healthy participants. In contrast to clinical neglect, however, the observed search deficits were observed uniformly across the search display. Although there are many general reasons why TMS might fail to mimic the behavioural changes caused by major brain lesions,this difference remains challenging, since TMS studies of line bisection have shown asymmetrical disruption (Bjoertomt, Cowey, & Walsh, 2002; Ellison et al., 2004). * Correspondence should be addressed to Professor A. D. Milner, Cognitive Neuroscience Research Unit, Wolfson Research Institute, University of Durham, Stockton-on-Tees, TS17 6BH, UK ( a.d.milner@durham.ac.uk). DOI: / X227024

2 502 I. Schindler et al. In the present experiments, we used the same search criteria and the same sites of TMS as Ellison et al. (2004). However, instead of simply detecting the presence or absence of a target stimulus, participants were asked to indicate the location of the target, i.e. in the left vs. right half of the display. In addition, we examined the effect of presenting small structured search arrays in different parts of the visual field, as opposed to the previous method of presenting full-field arrays. For comparability, we used the same subject cohort in these experiments as in our previous study, and we also restricted our investigation to the tasks forwhich specialization was found in each area, namely, difficult feature search for right STG and difficult conjunction search for right PPC. Methods Subjects Five healthy participants, aged 21 36, with normal or corrected to normal vision participated throughout. Participants gave informed consent in accordance with the Declaration of Helsinki and Durham University Ethics Advisory Committee. Participant selection complied with current guidelines for repetitive TMS research (Wasserman, 1998). TMS Stimulation was applied to the right hemisphere using a Magstime Rapid at 65% of the stimulator s maximum power. Localization and stimulation procedures directly replicated our previous experiments (Ellison et al., 2004; Figure 1) with a 50mm figure-of-8 coil delivering 4Hz stimulation over 500 ms to STG and a 70 mm figure-of-8 coil delivering 10 Hz stimulation over 500 ms to PPC. At each site, the train of pulses began at presentation of the stimulus array. Each search task was tested in 8 12 trials with alternate TMS and no-tms blocks. The order of all tasks was randomized across participants. Testing required three sessions (one per week). In no-tms blocks, a nondischarging coil was held over the area of interest whilst another coil discharged in close proximity. Thus, the subjective experience was the same as TMS blocks, but no magnetic pulse was administered (sham condition). Procedure The basic experimental setup replicated that of Ellison et al. (2004). Participants were asked either to make a present/absent decision or a left/right decision on each trial. The left button was used for target-present or target-left responses, the right button for target-absent or target-right responses. Each trial was preceded by a central fixation cross ( ) for 500 ms, followed immediately by the stimulus array. The array remained present until response or for 1,500 ms, whichever was shorter. The inter-trial interval was 4,000 ms. Experiment 1. Two visual search tasks were used (Figure 2). Both of the tasks (feature and conjunction) required serial search (. 10 ms/item) among eight-item and thus constitute hard search tasks in terms of difficulty. The target, which was present on every trial, could appear anywhere in a6 6array of virtual boxes (overall size of visual angle) on the monitor screen, equally often on the left or right. Stimulus items each subtended visual angle and were presented against a black

3 Contralateral visual search deficits following TMS 503 Figure 1. TMS stimulation sites. Stimulated areas were localized using each participant s MRI scan co-registered to their skull coordinates using BrainSighte software. background, with four always in the left half and four in the right half of the array. In the feature task, the target was entirely unique whereas in the conjunction task, both the orientation and colour of the target were present among the distractors (Figure 2). Participants were asked to respond as quickly and accurately as possible to indicate the position of the target. Experiment 2. A structured search array consisting of an eight-item square with each item subtending 18 the spatial location of the array was used, with either a target-present/absent response or a left/right response being required. The search array

4 504 I. Schindler et al. Figure 2. Visual search arrays used in Experiment 1. The target appeared equally on the left and right of the screen. Left, hard feature search; right, hard conjunction search. was presented either centrally, to the left or to the right (Figure 3), thus allowing comparison of performance either in relation to the array or in relation to space. The outer border of the lateral arrays extended 188 from the centre of the screen. The targets were presented in pseudo-random order at one of the six lateral positions within the square. During the present/absent response condition, the target was present on only 50% of trials. In the left/right response condition, the target was always present, occurring equifrequently on the left or right of the array. Analyses The data from each experiment were analysed in two steps. First, the raw reaction time (RT) data were subjected to repeated measures ANOVA, with post hoc comparisons using Bonferroni-corrected two-tailed t tests. Wherever TMS had a significant effect, additional analyses were performed on the relative TMS effect. This was done using TMS data normalized with respect to the relevant no-tms (sham) data: Normalized RT ¼ ðtms 2 shamþ*100=sham Figure 3. Structured search arrays used in Experiment 2. The search arrays were presented pseudorandomly on the left, centre or right of the screen. Top, typical arrays used in the hard feature task and bottom, typical arrays used in the conjunction task. Identical stimulus elements were present in the random full-field arrays used in Experiment 1 (see Ellison et al., 2004).

5 Results Contralateral visual search deficits following TMS 505 Error rates in all experiments were less than 3.5% and there were no significant differences between TMS and no-tms error rates. In all cases, data points lying outside two standard deviations from the mean in each condition were excluded. Experiment 1: Full search array, left/right responding ANOVAs on the raw RTs revealed a significant interaction between stimulation and hemispace [Fð 1 ; 4Þ¼37:21, p, :005] for PPC stimulation (conjunction search), with TMS-increased RTs for targets in the left [ tð 4 Þ¼4 : 29, p, : 05] but not the right hemispace [t ð4þ. 0 : 5; Figure 3, left].the TMS effectsdiffered significantly between left and right [ t ð 4 Þ¼3 : 36, p, : 05]. In the case of right STG stimulation (feature search), there was asignificant main effect of TMS [ F ð 1 ; 4 Þ¼9 : 81, p, : 05] but no significant effect involving hemispace (Figure 4). Experiment 2: Structured arrays presented centrally or laterally The raw data for Experiment 2 are shown in Table 1. In the detection condition, ANOVAs(stimulation array position target side) revealed a similar pattern for both PPC and STG stimulation: in both cases there was a significant main effect of array position [F ð 2 ; 8 Þ¼19: 27, p, : 001 for PPC; F ð 2 ; 8 Þ¼14: 09, p, : 01 for STG] with faster responses being made to central than to left or right arrays (means ms vs ms and ms for PPC, and ms vs ms and ms for STG). There was also a significant interaction between array position and side [ F ð 2 ; 8Þ¼11:26, p, : 005 forppc; F ð 2 ; 8 Þ¼26: 81, p, : 001 forstg], reflecting an overall tendency for the outer locations (left for the left array and right for the right array) to be searched slowest. However, there was no significant overall TMS effect, nor any interaction of TMS with array position or side of target presentation, in either case. In the left/right response condition with PPC stimulation, ANOVA showed a similar pattern to that seen for target present/absent responses, with a significant main effect for array position and a significant array position side interaction [ F ð2 ; 8Þ $ 8 :58, Figure 4. TMS effects on visual search. The data presented are percentage changes in reaction time caused by TMS. Left, the effects of posterior parietal stimulation during aconjunction search task using full-field array, with a left/right choice response. Right, the effects of superior temporal stimulation during a hard feature search task using a structured array located in different spatial positions, again when a left/right choice response was required.

6 506 I. Schindler et al. Table 1. Results of Experiment 2 No TMS Left array Central array Right array Stimulation area Response Left Right Left Right Left Right PPC PA (42.64) (40.37) (71.01) (33.00) (63.21) (50.92) LR (23.38) (31.76) (36.70) (18.65) (31.16) (14.51) STG PA (26.03) (68.24) (21.59) (49.23) (59.04) (50.94) LR (45.96) (71.44) (30.18) (78.99) (56.32) (81.40) TMS PPC PA (51.01) (57.25) (52.16) (47.39) (50.68) (36.67) LR (50.99) (37.99) (43.76) (51.80) (51.89) (39.04) STG PA (37.27) (57.81) (46.88) (37.53) (42.97) (42.47) LR (81.60) (52.37) (41.65) (71.54) (80.17) (71.86) Search data are presented as mean reaction times. Control (no-tms) data are given at the top and TMS data at the bottom. PPC, posterior parietal cortex (conjunction search); STG, superior temporal gyrus (feature search). PA, present/absent judgements; LR, left/right judgements. Small structured search arrays were presented either in left hemispace, right hemispace or centrally. SEs are given in brackets.

7 Contralateral visual search deficits following TMS 507 p, : 05]. TMS had no significant effect. In contrast, asimilar analysis of the STG data revealed significant main effects for TMS and array position [ F ð 1 ; 4 Þ $ 8 : 73, p, : 05], the latter reflecting much faster RTs for central arrays (mean ms) than for left (785.7 ms) or right (756.0 ms) arrays. There were also significant interactions between position side, stimulation side and stimulation position side [ F ð 1 ; 4 Þ $ 12: 82, p, : 05]. Post hoc t tests revealed asignificant RT increase with TMS for targets on the left side of an array, when itappeared in either left or right hemispace [(t ð 4 Þ $ 5 : 38, p, : 01],though not whenitappeared centrally [ t ð 4 Þ¼2 : 34, p. : 05]. Analyses of the normalized reaction times at each array position separately confirmed greater left-side than right-side TMS effects for left and right arrays [ t ð 4 Þ $ 7 : 22, p, : 01], but not central arrays [ t ð 4 Þ¼1 : 72, p. : 2]. Discussion We have extended our previous findings that TMS over right STG causes deficits in processing during single feature search, while TMS over right PPC adversely affects processing during conjunction search, by demonstrating that under suitable conditions, TMS to these areas can yield searchdeficitsrestricted to the contralateral side. Thus, we have found, for the first time, biased searchdeficits following TMS, reminiscent of those associated with unilateral visual neglect. Only when the response required a left/right decision on target location, however,were these contralateral effects of TMS uncovered. Evidently their appearance was contingent on the explicit (not just the incidental) processing of target location. Our intention was that the location task would force participants to code stimulus items not only in terms of their identity, but simultaneously also in terms of their location, as would typically be the case in everyday behaviour where the aim is to find and then respond to atarget object. In our view, aperceptual binding between object and location would thus be needed only in the location task. However, there may have been other differences in cognitive processing that could have contributed to the results. For example, although the ultimate orienting of gaze towards the target was necessary for either task to be performed successfully, only in the location task would there be a spatial compatibility between this orienting and the correct manual choice response itself. This could have been afactor in producing the present patternof results, particularly those caused by PPC stimulation. This account would be less persuasive in the case of the STG results, however,in that the asymmetrical results there were present both in the left and the right side of visual space. Anovel finding of the present experiments lies in the different ways in which the right PPC and STG seem to be engaged in coding visual space. By manipulating the spatial location of the search array, wefound that right STG stimulation caused adeficit when targets appear on the left side of an array, whichever side of space the array was located. In other words, the PPC seems to be involved contralaterally in the broad processing visual space, whereas the STG is involved contralaterally only in respect of the search array itself. This array-based involvement of the right STG is reminiscent of reports of object-based neglect in some neglect patients (Walker, 1996). It remains at first sight apuzzle as to why this effect did not appear with central presentations of the search arrays. Onepossible reason is that there was akind of ceiling effect, since search times in Experiment 2 were much quicker for central than for peripherally presented arrays. There is of course a general tendency for any visual task to

8 508 I. Schindler et al. be performed more efficiently in central vision, and indeed many cortical visual areas contain a neural over-representation of central stimuli. This is probably true of the STG area we are studying here. In parenthesis, it should be noted that the lack of a TMS main effect in the detection condition of Experiment 2 does not contradict our previous report (Ellison et al., 2004). In deliberate contrast to that study, here the arrays were smaller and were structured as a geometrical Gestalt, thereby enabling participants to perceive the entire array within a single fixation. Our findings may be taken as further evidence of the different processing roles of the right PPC and right STG: not only in terms of task specificity but also in terms of different frameworks for spatial processing in the normal brain. They suggest the testable possibility that a patient may show signs of unilateral neglect differentially in neuropsychological testing not only in a task-related manner, but in a spatially specific manner, depending on the location of their brain damage. References Bjoertomt, O., Cowey,A., &Walsh, V. (2002). Spatial neglect in near and far space investigated by repetitive transcranial magnetic stimulation. Brain, 125, Ellison, A., Schindler, I., Pattison, L. L., & Milner, A. D.(2004). An exploration of the role of the superior temporal gyrus in visual search and spatial perception using TMS. Brain, 127, Gharabaghi, A., Fruhmann Berger, M., Tatagiba, M., &Karnath, H.-O. (2006). The role of the right superior temporal gyrus in visual search Insights from intraoperative electrical stimulation. Neuropsychologia, 44, Karnath, H.-O., Ferber, S., & Himmelbach, M. (2001). Spatial awareness is a function of the temporal not the posterior parietal lobe. Nature, 411, Karnath, H.-O., Fruhmann Berger, M., Kuker, W., & Rorden, C. (2004). The anatomy of spatial neglect based on voxelwise statistical analysis: A study of 140 patients. Cerebral Cortex, 14, Karnath, H.-O., Milner, A. D., & Vallar, G.(Eds.). (2002). Cognitive and neural bases of spatial neglect. Oxford: Oxford University Press. Mort, D. J., Molhotra, P.,Mannan, S. K., Rorden, C., Pambakian, A., Kennard, C., et al. (2003). The anatomy of visual neglect. Brain, 126, Vallar, G., & Perani, D. (1986). The anatomy of unilateral neglect after right-hemisphere stroke lesions. A clinical/ct-scan correlation study in man. Neuropsychologia, 24, Walker, R.(1996). Spatial and object-based neglect. Neurocase, 1, Wasserman, E. M. (1998). Risk and safety of repetitive transcranial magnetic stimulation: Report and suggested guidelines from the International workshop on the safety of Repetitive Transcranial Magnetic Stimulation, June 5 7, Electroencephalography and Clinical Neurophysiology, 108, 1 16.