Electrophysiological abnormalities of spatial attention

Objective: We evaluated event-related potentials (ERPs) elicited by attentional disengagement in individuals with autism. Methods: Sixteen adults with autism, 17 adults with mental retardation and 14 healthy adults participated in this study. We recorded the pre-saccade positive ERPs during the gap/overlap task under which a peripheral stimulus was presented subsequent to a stimulus in the central visual field. Under the overlap condition, the central stimulus remained during the presentation of the peripheral stimulus and therefore participants need to disengage their attention intentionally in order to execute the saccade to the peripheral stimulus due to the preservation of the central stimulus. Results: The autism group elicited significantly higher pre-saccadic positivity during a period of 100 to 70 msec prior to the saccade onset than the other groups only under the overlap condition. The higher amplitude of pre-saccadic positivity in the overlap condition was significantly correlated with more severe clinical symptoms within the autism group. Conclusions: These results demonstrate electrophysiological abnormalities of disengagement during visuospatial attention in adults with autism which cannot be attributed to their IQs. Significance: We suggest that adults with autism have deficits in attentional disengagement and the physiological substrates underlying deficits in autism and mental retardation are different. 3 Introduction Autism is diagnosed on the basis of behavioral and developmental features such as impairment of reciprocal social interaction, communication and imagination, and the presence of repetitive and ritualistic behavior. It has been recently pointed out that scientific merits from studies using a comprehensive model assuming a single explanation for autism may be limited, since various behavioral features of autism are largely independent in terms of genetic background as shown by twin cohort studies (Happé et al., 2006). Thus, a bottom-up type approach which explores abnormal or preserved functions in relatively fundamental cognitive processes based on simple cognitive models may be important in the autism research strategy. Researchers have proposed that attentional abnormalities underlie the behavioral features of autism such as inflexibility, repetitive behavior and overselectivity (Lovaas et al., 1979; Casey et al., 1993; Townsend et al., 1996,2001; Wainwright and Bryson, 1996, Senju et al., 2004). Auditory attention has been intensively studied in the literature, particularly using event-related potentials (ERPs). For example, a reduced P300 amplitude (Oades et al., 1988; Lincoln et al., 1993) and abnormal mismatch negativity (Gomot et al., 2002; Kasai et al., 2005) has been reported. On the other hand, spatial attention has been also implicated in the pathophysiology of autism, which has been understudied. Spatial attention is a function of spatially directed attention and spatial selectivity (Allport, 4 1989). Posner and Cohen (1984) outlined a model in which the successive components of spatial attention were defined as disengagement, shift, and engagement of attentional sources. In Posner’s cuing paradigm, a target stimulus follows a cue stimulus presented on the same side of a target (valid condition) or on the contralateral side of a target (invalid condition). This task makes it possible to investigate the ability to disengage attention from one location and shift attention to its contralateral location. Findings from patients with acquired brain damage suggest that these components are associated with specific brain areas, i.e., disengagement is associated with the parietal cortex, shift is related with superior colliculus, and engagement is associated with thalamus (Posner and Cohen, 1984; Posner and Petersen, 1990). Most studies on visuospatial attention in adults with autism have used the cuing paradigm that elicits the re-orienting of the direction of the attention (Posner, 1980) and reported behavioral problems in attentional disengagement and shift (Casey et al., 1993; Townsend et al., 1996,2001; Wainwright and Bryson, 1996), but findings of other studies using eye gaze cue have been mixed (Senju et al., 2004; Swettenham et al., 2003; Kylliäien and Hietanen, 2004). Other studies reported that autistic children presented the difficulties disengaging attention from the salient object (Hughes & Russell, 1993) and mental concepts (Hughes et al, 1994). On the other hand, studies using the gap paradigm (Saslow, 1967) have shown impairment of attentional engagement in children (van der Geest et al., 2001) and adults (Kawakubo et al., 2004) with autism. In the 5 gap/overlap task, when a temporal gap is introduced between the disappearance of a central fixation point and the appearance of a new target stimulus, the saccade reaction times are reduced compared to when no gap is introduced (gap effect; Saslow, 1967). This difference in saccade reaction times has been explained by the difference in attentional disengagement(Fischer and Weber, 1993). In the gap condition, in which an initial fixation point disappears before a target appears, attention on the fixation point is disengaged automatically. On the other hand, in the overlap condition, in which the fixation point remains after the target appears, attention on the fixation point is disengaged due to the appearance of the peripheral target stimulus. Therefore, the saccade reaction times in the overlap condition are longer than in the gap condition. It cannot be concluded, however, that individuals with autism show only impairment of attentional engagement but not of disengagement in the gap paradigm. If the engaging of attention to the central fixation point is enhanced, they may show impairment of attentional disengagement. In order to clarify whether individuals with autism have an impairment of attentional disengagement in the gap paradigm, further examination under conditions in which attentional engagement to the central visual field is ensured is necessary. Moreover, little is known of the physiological substrates for these visuospatial attentional deficits in autism. In the visual ERP literature, P300 amplitude of individuals with autism under a condition where the target stimulus was presented in the center of visual field was comparable to 6 that of normal controls (Pritchard et al., 1987; Ciesielski et al., 1990). On the other hand, when the stimulus was presented at various peripheral locations individuals with autism were associated with smaller-than-normal visual P300 amplitude (Kemner et al., 1999; Townsend et al., 2001). To our knowledge, however, no previous studies have directly examined ERPs elicited by attention disengagement in individuals with autism. Csibra et al. (1997) and Gómez et al. (1996) analyzed saccade-locked ERP during the gap/ overlap task in healthy adults, and found a slowly developing pre-saccadic positivity elicited at central-parietal electrode sites that was followed by a pre-saccadic spike potential immediately preceding the eye movement. Csibra et al. (1997) concluded that attentional disengagement in the gap/overlap task was reflected in these parietal positive ERP components. Thus, the goal of this study was to assess the pre-saccade positive ERP components during the gap/overlap task as an electrophysiological index of attentional disengagement in adults with autism. To ensure attentional engagement to the central visual field, our task was different from a typical format of the gap/overlap task in some points that various pictures were used as stimuli instead of a cross and a square and participants were required to discriminate the stimuli. Moreover, to clarify whether the physiological dysfunction of visuospatial attention was specific to autism rather than being attributable to general intellectual disability, we compared the data from adults with autism with those of IQ-matched adults with mental retardation. Previous 7 studies (Csibra et al., 1997, Gomez et al., 1996) indicated that the pre-saccadic positivity was higher in the overlap condition than in the gap condition. Thus, it is assumed that the pre-saccadic positivity reflects the resource that is needed to disengage the attention. Accordingly, we hypothesized that the autistic group would exhibit impairment in attentional disengagement under the overlap condition that would be reflected as higher or longer pre-saccadic positivity detected in the saccade-locked ERPs. Moreover, we predicted higher presaccadic positivity would be associated with severer autistic symptoms. Method Participants Sixteen adults with autism (11 men and 5 women; mean age [SD], 29.0 [6.5]; mean IQ [SD], 43.6 [14.7]), 17 adults with mental retardation (12 men and 5 women; mean age [SD], 27.5 [5.0]; mean IQ [SD], 40.6 [10.9]), and 14 healthy adults (6 men and 8 women; mean age [SD], 28.5 [5.8]; mean IQ [SD], 105.2 [10.9]) participated in this study. Age (one-way ANOVA, p=0.70) and gender (chi-square test, p=0.23) were not significantly different among groups. IQs were evaluated with the Wechsler Adult Intelligence Scale-Revises (WAIS-R) for two of autism participants and with the Tanaka-Binet Test for eleven of autism participants and all individuals with mental retardation. For healthy adults IQs were estimated by four subtests of the WAIS-R. 8 Autism subjects and mentally-retarded subjects did not significantly differ in IQ (t-test, p=0.52). Diagnosis of autism was determined according to DSM-IV criteria and Childhood Autism Rating Scale (CARS: Schopler et al., 1988) by a trained pediatric psychiatrist (clinical experience > 5 years; KW, HK, or MI). We used a CARS score of 27 as the cutoff point (Mesibov et al., 1989). Autism subjects had significantly higher CARS scores than mentally-retarded subjects (34.1 [SD=4.1] versus 20.3 [3.1]; t-test, p<0.001). All participants had normal or corrected to normal vision and did not have any chromosomal or neurological disorders. All participants were right-handed determined using the Edinburgh Inventory. The ethical committee of the Faculty of Medicine, University of Tokyo, approved this study (No. 629). Written informed consent was obtained from all participants and their parents before the experiment. Stimuli and Apparatus The stimuli were presented on a 21-inch CRT display with a black background by using STIM software (Neuroscan, Inc.). To encourage the motivation of the participants, we used as stimuli 13 illustrations that consisted of pictures of animals (e.g., dog, frog), daily goods (e.g., glasses, baggage) and vehicles (e.g., bicycle, car), etc. In place of a fixation point a central stimulus (picture stimulus as above) was presented in the central visual field. Peripheral stimuli were presented 14 degree (deg.) to the left or right of the central stimulus. All stimuli were the 9 same angular size; 1.3 X 1.3 deg. Please insert Figure 1 about here. Procedure Participants were seated 0.60 m away from the CRT monitor with their chin on a chinrest. Participants initiated each trial by pressing a start button, and after 1000 msec a central stimulus was presented in the center of the CRT monitor. A peripheral stimulus was then presented to the left or right side of the central stimulus for 2000 msec. To minimize the possibility that participants anticipate the timing of the peripheral stimulus onset, the central stimulus was presented at intervals randomly varying between 700 and 1500 msec. To make the task clearer to participants, we instructed them to press a button when a target stimulus, the drawing of a dog, appeared as either the central stimulus or the peripheral stimulus. In some trials, the dog was presented both in the center and the peripheral visual field. Thus, even if the dog was presented in the center, saccades were required. Since it was difficult to ask button pressing at an appropriate timing for three subjects, these subjects were alternatively asked to speak aloud the name of the target stimulus. It was necessary for participants to move their eyes to the peripheral stimulus when it was presented, in order to discriminate the target stimulus from the other 10 stimuli. In the gap condition, the central stimulus disappeared 200 msec before the peripheral stimulus was presented, and in the overlap condition, the central stimulus remained during the presentation of the peripheral stimulus. Practice trials were given to each participant to ensure that the instructions had been understood. After the practice trials, 180 trials composed of 45 trials of each of the four conditions ([Gap vs. overlap] X [leftvs. right-side presentation on the monitor]) were presented in a random order in three blocks of 60 trials each. The probability of the appearance of target stimulus was 20%. Recordings The scalp electroencephalogram (EEG) was recorded according to the international 10-20 system using Ag/AgCl electrode caps (Neuroscan, Inc.) at 16 electrode sites (F3, Fz, F4, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, O1, Oz, O2) referred to linked earlobes. The horizontal electrooculogram (EOG) was recorded at the outer canthi of the both eyes and the vertical EOG was recorded from electrodes placed below and above the left eye. The bandpass filter was set at 0.15-30 Hz and the sampling rate was 500 Hz. EEGs and EOGs were analyzed using SCAN system with SynAmps (Neuroscan, Inc.).