Autism spectrum disorder (ASD) is primarily characterized by impairments in social skills and by restricted and repetitive patterns of behavior (American Psychiatric Association, 2013). Although increased or decreased responsiveness to sensory input is also part of the current diagnostic criteria, the nature of atypical perceptual function in ASD is not fully understood, and not part of the diagnostic criteria. Furthermore, the connection between altered perceptual function and the core behavioral symptoms of ASD is unclear.

Classic theories suggest that atypical perceptual function in ASD reflects a shift in favor of processing local (or low-level) sensory details, over integration of those details into a global (or high-level) percept. This may reflect (non-exclusively): impaired global processing (Shah and Frith, 1983; Frith and Happé, 1994; Behrmann et al., 2006; Booth and Happé, 2018), superior local processing (Plaisted et al., 2003; Mottron et al., 2006; Mottron and Burack, 2021), and/or a preference for detail (Plaisted et al., 1999; Happé and Frith, 2006; Koldewyn et al., 2013). In recent years, it has been proposed that the bias toward low-level processing in ASD, as well as other perceptual and sensory atypicalities, such as hyper- or hypo-sensitivity to sensory stimuli (Rogers and Ozonoff, 2005), may be explained by altered perceptual inference (Pellicano and Burr, 2012; Friston et al., 2013; Lawson et al., 2014).

Perceptual inference has been broadly studied in typical brain function within the Bayesian framework (Knill and Richards, 1996; Kersten and Yuille, 2003; Knill and Pouget, 2004; Mamassian et al., 2018). The Bayesian approach provides a principled (computational) account of perception as a process of statistical inference, and posits that top-down expectations (‘priors’) are integrated together with bottom-up sensory signals (which are inherently ‘noisy’, that is, variable) to improve perceptual reliability. Although integrating ‘priors’ reduces sensory noise in favor of high-level (global) perception, this might be at the expense of low-level sensory accuracy. Accordingly, Pellicano and Burr, 2012 proposed that individuals with ASD might use priors to a lesser degree, resulting in superior local processing and altered sensory responsiveness. However, relatively greater sensory (vs. prior) weighting might actually be expected from Bayesian theory due to superior low-level sensory function in ASD (Brock, 2012; Karvelis et al., 2018). Thus, whether or not priors are altered and what aspects of perceptual inference may differ in ASD requires investigation.

Priors are complex – they reflect a compound (subjective) aggregate of an individual’s beliefs about the probability of observing various events or stimuli in the world (Ma, 2019). This is influenced by context, statistics of the environment, and other factors, across broad and narrow timescales. For example, one’s prior for hearing explosions might temporarily increase when watching an action movie (but return to normal when leaving the cinema). By contrast, the prior that objects fall to the ground (due to gravity) might be more long-term and constant. It is unclear therefore what aspects of priors might be affected in ASD. At minimum, these can be divided according to their recency: short-term (recent information such as context or immediately preceding events) and long-term (such as beliefs and expectations about the world that were learned throughout life).

Recent studies have found that long-term priors seem intact in ASD. The ‘light-from-above prior’ (that light more frequently comes from above) was applied by individuals with ASD (Croydon et al., 2017), the influence of object size and brightness on perceptions of heaviness were similar in ASD (Buckingham et al., 2016; Hadad and Schwartz, 2019), and a slow-speed prior was integrated in ASD (Noel et al., 2020) – like controls. These results indicate that sensory predictions based on statistical information gathered over long-term experience are indeed built and integrated in ASD. Also, an experiment that tested ‘Mooney’ images (difficult to interpret pictures, comprising black and white patches) showed that prior exposure to the high-fidelity images helped individuals with ASD as much as controls (Van de Cruys et al., 2018), indicating that priors over a time-course of minutes are also used in ASD. Thus, an overall reduction of all priors in ASD does not seem likely.

By contrast, studies that found reduced adaptation to prior stimuli in ASD seem to provide evidence for a reduced impact of recent sensory information (in the range of seconds) on perception in ASD. For example: prolonged exposure to a specific facial identity, which typically biases subsequent perception away from that image, did so to a lesser degree in ASD (Pellicano et al., 2007). Also, adaptation to numerosity judgements (Turi et al., 2015), to loudness (Lawson et al., 2015), and to audio-visual asynchronies (Noel et al., 2017) was reduced in ASD. At face value, these results seem to support the notion of reduced use of recent prior information in ASD. However, ‘adaptation’ (like in these studies) most commonly refers to the phenomenon by which exposure to a stimulus repels subsequent perception away from that stimulus (often called a ‘negative’ effect) – opposite to a Bayesian prior, which generally attracts perception towards the more frequent stimuli (positive effect; although Bayesian priors can also lead to negative effects, for example Peters et al., 2016). Thus, while reduced adaptation in ASD may suggest a smaller impact of prior stimuli, this does not imply reduced priors. Also, adaptation is an umbrella term that includes a repertoire of effects (Kohn, 2007) – what aspects thereof are specifically altered in ASD requires investigation.

‘Serial dependence’, a related topic of high recent interest, also studies the impact of recent sensory information (in the range of seconds) on subsequent perception. Specifically, it describes a positive effect by which perceptual decisions are biased toward recently experienced stimuli (Fischer and Whitney, 2014; Liberman et al., 2014; Taubert et al., 2016; Alexi et al., 2018; Liberman et al., 2018) – opposite to the negative effect of adaptation. Here, too initial evidence shows reduced serial dependence in ASD (Molesworth et al., 2015; Lieder et al., 2019), suggesting a smaller impact of recent stimuli in ASD.

However, these studies of reduced serial dependence and reduced adaptation in ASD only take into account the previous stimuli, and not the previous choices. This is important because there is mounting evidence that prior perceptual decisions (rather than sensory stimuli) elicit the positive effect of serial dependence (Kaneko and Sakai, 2015; Talluri et al., 2018; Feigin et al., 2021). Thus, the overall impact of recent sensory experience might reflect a composite effect, for example of repulsion away from the prior stimuli (adaptation) and attraction toward the prior choices (serial dependence) which can, at least partially, cancel one-another (Fritsche et al., 2017; Pascucci et al., 2019). Therefore, understanding specifically which response (to recent prior stimuli or prior choices) is altered in ASD may change our understanding of these results profoundly.

The effect of prior choices is a factor that has been less considered so far in ASD. A recent study of high-level (executive) decision-making found greater consistency in the decisions of individuals with ASD, which may suggest an ASD tendency toward repetitive behavior also in decision-making (Wu et al., 2018). However, this study used a gambling task, which may be different from perceptual decision-making, and thus did not dissociate the serial effects of prior sensory from prior choice information, which needs to be tested together in a perceptual decision-making task. We hypothesized, based on the studies described above, that a positive effect of serial dependence from prior choices may be enlarged in ASD, and that this may cancel out adaptation effects, which could appear as reduced adaptation in ASD if the effects of prior stimuli and choices are not accounted for separately.

In this study, we aimed to dissociate how recent prior stimuli and choices (over a time-course of seconds) affect perceptual decisions in ASD. For this purpose, we tested the performance of individuals with ASD and controls in a visual location discrimination task designed to induce short-term biases over several trials (Feigin et al., 2021). In addition, we reanalyzed data from a previous study of multisensory (visual-vestibular) heading direction discrimination performed on a different group of participants (Zaidel et al., 2015). Both tasks comprised a two alternative forced choice (2AFC) regarding a single interval stimulus. For the location discrimination task, participants discriminated whether a visual stimulus (solid circle, presented on a screen) lay to the left or right of the screen center. For heading discrimination, participants discriminated whether a linear self-motion stimulus was leftward or rightward of straight ahead. We fit the data with a logistic regression model that separated between the effects of prior stimuli and prior choices.

Strikingly, in both tasks (two different cohorts), we found a stronger effect of prior choices on perceptual decisions in ASD vs. controls. By contrast, individuals with ASD demonstrated a similar effect of adaptation to previous stimuli compared to controls (in both tasks). These results challenge theories of reduced priors in ASD, and rather implicate an increased influence of prior choices in ASD. This may expose a novel aspect of repetitive behavior in ASD, reflected in perceptual decisions.

The main aim of this study was to investigate whether the influence of recent prior information on perceptual decision-making is altered in ASD. This was tested on two different datasets: (i) visual location discrimination (collected for this study), and (ii) multisensory visual-vestibular heading discrimination (reanalysis of previously published data, Zaidel et al., 2015). In both datasets, we found a stronger effect of recent prior information on perceptual decisions in ASD compared to controls. Moreover, we found that this effect was mediated by a greater influence of prior choices in ASD. Specifically, perceptual decisions were more likely to be the same as previous choices (an increased ‘consistency’ bias). Interestingly, no differences were found regarding the influence of recent sensory information.

Location discrimination

In the visual location discrimination task, the participants were required to report whether the visual stimulus (a filled circle) lay to the right or to the left of the screen center (2AFC). Stimulus locations were biased to the right or to the left for several ‘prior’ discriminations, followed by a single unbiased (i.e. balanced) test stimulus (Figure 1; rightward and leftward biased trials were interleaved). We first tested the overall effect of these short-term priors by sorting the test stimuli according to their prior bias (right/left) and plotting separate psychometric functions for each prior type (per participant).

Figure 1

Download asset Open asset

Larger effect of recent priors in ASD for location discrimination

Example psychometric plots for one participant with ASD and one control participant (red and blue shades, left and right subplots of Figure 2A, respectively) reveal that for both participants test stimulus discriminations were influenced by recent priors. This is seen by a shift in the psychometric curves for rightward vs. leftward biased priors (light vs. dark shades, respectively). Note that all psychometric curves are plotted as a function of the test stimulus location (which was unbiased). Thus, any differences between the light- and dark-shaded curves reflect different responses to the same test stimuli (therefore shifts in the psychometric curves expose the effects of recent priors).

Figure 2 with 1 supplement see all

Download asset Open asset

To understand the direction of the shift – it is easiest to compare the psychometric curves where they cross x = 0° (vertical dotted line). For both participants, the lighter shaded curve (right biased priors) crosses at a higher value vs. the darker curve (left biased priors). This means that there is a higher probability for choosing right following rightward biased priors (and a higher probability for choosing left following leftward biased priors). Another way to view and to quantify this is by the difference between the curves’ PSEs (where they cross the horizontal dotted line y = 0.5). A larger PSE difference (ΔPSE, see Equation 4 in the Materials and methods) indicates a stronger effect, and positive ΔPSE values (like these examples) mean that the participant’s choices were biased in the direction of the recent priors. In these examples, the ΔPSE was larger for the participant with ASD (difference between the red-shaded curves) versus the control participant (difference between the blue-shaded curves).

At the group level, the ΔPSEs were highly significant for the ASD group (red bar, Figure 2B; t₁₇ = 5.2, p = 6.7∙10⁻⁵, Cohen’s d = 1.23, 95% CI = [0.6, 1.84], t-test), and smaller, but still (albeit marginally) significant for the control group (blue bar, Figure 2B; t₁₉ = 2.1, p = 0.047 Cohen’s d = 0.47, 95% CI = [0.006, 0.93], t-test). Notably, the ΔPSEs were significantly larger for the ASD vs. control groups (t₃₆ = 2.1, p = 0.043, Cohen’s d = 0.68, 95% CI = [0.02, 1.3], t-test). Hence, not only was a significant effect of recent priors seen also in ASD, it was larger than that of controls. This result (and the results presented below) does not result from a difference in response times between the groups, because these were similar: mean ± SD = 0.73 ± 0.11 s for ASD and 0.78 ± 0.13 s for controls (calculated across all discriminations, that is, pooling prior and test discriminations; t₃₆ = −1.3, p = 0.2, Cohen’s d = 0.42, 95% CI = [−1.06, 0.23], t-test).

Enhanced influence of recent choices in ASD for location discrimination

The above (ΔPSE) analysis suggests that individuals with ASD have a larger influence of recent priors. However, it does not address the underlying reason – specifically, it does not dissociate the separate influences of prior stimuli from prior choices, which may be different. Recent studies have shown that the influence of recent prior stimuli and choices might be in opposite directions – that is, the effects or prior stimuli could be adaptive (reflected by a negative shift), whereas recent choices may lead to a consistency bias (Kaneko and Sakai, 2015; Talluri et al., 2018; Pascucci et al., 2019; Feigin et al., 2021). Thus, the effects of prior stimuli and choices could cancel each other out. Furthermore, a larger ΔPSE shift in ASD could result from a weaker stimulus adaptation effect, or a stronger influence of prior choices (or a combination thereof). Hence, conclusions cannot be drawn from the ΔPSE results alone.

To address this, we fit a logistic regression model (per participant) that predicted each participant’s discriminations of test stimuli, based on four factors: the current stimulus, previous stimuli, previous choices and a general subjective baseline bias (Figure 3). The fourth regression coefficient (baseline bias) was not part of the group comparison because this analysis aimed to investigate the difference in priors, and there was no expected difference in baseline bias (it was, however, still needed for the model fits in order to better estimate the other parameters). We confirmed that the baseline coefficients (β₀) were indeed not significantly different between the groups (t₃₆ = −1.46, p = 0.15, Cohen’s d = −0.47, 95% CI = [−1.1, 0.18], t-test). A comparison of the other three regression coefficients revealed a significant difference between the ASD and control groups (F_3,34 = 3.3,p = 0.03, η_p² = 0.23, one-way MANOVA).

Figure 3

Download asset Open asset

Further investigation into this group difference (to assess whether individuals with ASD were affected differently by the previous stimuli or the previous choices), using pairwise comparisons, found that the influence of previous choices (β_{prev_choices} coefficients) was significantly larger in ASD vs. controls (Figure 4, center plot; t₃₆ = 2.5, p = 0.034, Cohen’s d = 0.8, 97.5% CI = [0.05, 1.57], t-test, p-value reported here was multiplied by two to correct for multiple comparisons, see Statistics section in the Materials and methods for more details). Thus, while both groups were influenced by their previous choices in a ‘positive’ (attractive) manner (β_{prev_choices} coefficients were positive and significant for each group individually: t₁₇ = 9.9, p = 2∙10⁻⁸, Cohen’s d = 2.3, 95% CI = [1.42, 3.23] and t₁₉ = 6.4, p = 4∙10⁻⁶, Cohen’s d = 1.43, 95% CI = [0.8, 2] for ASD and controls, respectively; t-tests) this effect was significantly increased in ASD. This not only negates hypotheses of reduced priors in ASD – it shows the opposite – an increased influence of prior choices in ASD.

Figure 4 with 2 supplements see all

Download asset Open asset

By contrast, the influence of previous stimuli (β_{prev_stimuli} coefficients) was small (Figure 4, right plot; to enable comparison between choice and stimulus coefficients, stimulus values were normalized before model fitting to have RMS = 1, like choice). For both groups, the influence of previous stimuli tended to be negative, suggesting an adaptive effect. This was significant for the control group (t₁₉ = −2.5, p = 0.021, Cohen’s d = 0.57, 95% CI = [−1.03,–0.09], t-test) and close to significant for the ASD group (t₁₇ = −1.99, p = 0.063, Cohen’s d = 0.47, 95% CI = [−0.95, 0.03], t-test). There was no significant difference between groups (Figure 4; t₃₆ = 0.01,p = 0.99, Cohen’s d = 0.003, 95% CI = [−0.63, 0.64], t-test; the raw p-value is reported here before correction for multiple comparisons).

Lastly, as expected, the current stimulus coefficients (β_{curr_stimulus}) were large, positive and significant for both groups (Figure 4, left; t₁₇ = 9.2, p = 5∙10⁻⁸, Cohen’s d = 2.18, 95% CI = [1.3, 3.03] and t₁₉ = 8.1, p = 1.5∙10⁻⁷, Cohen’s d = 1.8, 95% CI = [1.08, 2.51] for the ASD and control groups, respectively; t-tests). Also, these did not differ significantly between groups (t₃₆ = -0.85, p = 0.4, Cohen’s d = 0.28, 95% CI = [-0.92, 0.37], t-test). This indicates that sensitivity did not differ between the groups. We confirmed this by comparing thresholds from the psychometric functions (see Materials and methods), which also did not differ significantly between groups (t₃₆ = 0.96, p = 0.34, Cohen’s d = 0.31, 95% CI = [-0.33, 0.95], t-test; BF₀₁ = 2.21). Hence, the only difference we found in ASD was an increased influence of previous choices. Lastly, stimulus sensitivity (${\beta }_{curr\_stimulus}$) did not correlate significantly with the prior choice effect ${\beta }_{prev\_choices}$ (r₃₆ = -0.18, p = 0.27).

Larger effect of recent priors in ASD for heading discrimination

To calculate ΔPSEs, we constructed psychometric curves conditioned on the previous trial. Based on our results from the location discrimination experiment (that recent choices drive the effect), the leftward or rightward prior type for each trial was defined according to the choice on the previous trial. We first present the ΔPSE results here, and in the following section (below) we present fits to the logistic regression model.

Psychometric curves for an example ASD participant and an example control participant (visual condition) are presented in Figure 5A (red- and blue-shaded curves, in the left and right plots, respectively). Here too, the psychometric curves were shifted, and in the same (positive) direction as the location discrimination results: namely, following rightward choices, test stimuli were more likely to be discriminated rightward, and following leftward choices, test stimuli were more likely to be discriminated leftward. Like above, in the examples here the shift for the ASD participant was larger (red colors) vs. the control participant (blue colors).

Figure 5 with 1 supplement see all

Download asset Open asset

We calculated ΔPSEs for each participant, per condition: visual, vestibular, and combined visual-vestibular cues (Figure 5B, left, right, and center plots, respectively). Since the data comprised several blocks with different levels of visual coherence, ΔPSEs were calculated per coherence (presented along the x-axis). Also in these data, the ΔPSEs were consistently larger for the ASD group vs. controls (the red lines lie above the blue lines in Figure 5B). This difference was statistically significant (F_1,24 = 4.74, p = 0.040, η_p² = 0.17, Mixed ANOVA). These results suggest that our findings of a greater influence of prior choices in ASD generalize across multiple sensory domains. But once again, further analysis was required in order to dissociate the effects of recent stimuli and choices.

Enhanced influence of recent choices in ASD for heading discrimination

To separate the influence of previous stimuli from previous choices, we fit the data using a logistic regression model, similar to location discrimination results presented above (Figure 3). But, here only one ‘prior’ was used (n = 1) because there was no consistent bias (by design) beyond the previous trial. In line with the location discrimination results, the heading discrimination model fits revealed a significant difference between ASD and control groups in the weight given to the previous choice (β_{prev_choices}; Figure 6, center; F_1,24 = 4.83, p = 0.038, η_p² = 0.17, Mixed ANOVA). The β_{prev_choices} coefficients were significantly positive for the ASD group (t₁₂ = 3.03, p = 0.010, Cohen’s d = 0.84, 95% CI = [0.06, 0.40], t-test), indicating that perceptual choices were attracted towards (i.e. more likely to be similar to) the previous choice. By contrast, the β_{prev_choices} coefficients were not significant in the control group (t₂₁ = −0.01, p = 0.99, Cohen’s d = 0.003, 95% CI = [−0.2, 0.2], t-test). Thus, here too, individuals with ASD demonstrated a greater influence of prior choices.

Figure 6

Download asset Open asset

The previous stimulus (β_{prev_stimuli}) coefficients were significantly negative for both the ASD and control groups (Figure 6, right plot; t₁₃ = −6.73, p = 1.4∙10⁻⁵, Cohen’s d = 1.8, 95% CI = [−0.62,–0.32] and t₂₁ = −6.83, p = 9.5∙10⁻⁷, Cohen’s d = 1.46, 95% CI = [−0.51,–0.27], respectively; t-tests). This indicates a significant (negative) effect of adaptation to previous stimuli in the heading discrimination experiment. Accordingly, adaptation seemed to be stronger in heading discrimination vs. location discrimination, which showed a similar trend but with marginal statistics (possibly due to the different durations of the stimuli: 1 s for heading stimuli and 0.25 s for location stimuli). Notably, here too there was no significant difference between the β_{prev_stimuli} coefficients for the ASD and control groups (F_1,23 = 0.74, p = 0.40, η_p² = 0.03, Mixed ANOVA). Thus, we found no evidence for reduced adaptation in ASD. Rather, we found an increased (positive) influence of prior choices. Since these effects can be in opposite directions, if they aren’t dissociated, a greater influence of prior choices could (mistakenly) be interpreted as reduced adaptation. For example, in our vestibular data, the overall (ΔPSE) shifts lie closer to zero for ASD vs. controls (Figure 5B, right plot). However, this does not reflect reduced adaptation, but rather a greater influence of prior choices in ASD.

Naturally, current stimulus coefficients (β_{curr_stimulus}) were large, positive, and significant in both ASD and control groups (Figure 6, left; t₁₂ = 13.35, p = 6.8∙10⁻⁸, Cohen’s d = 3.7, 95% CI = [3.8, 5.29] and t₂₁ = 15.7, p = 4.46∙10⁻¹³, Cohen’s d = 3.35, 95% CI = [4.41, 5.76], respectively; t-tests), indicating that, like in the location discrimination task, the current stimulus was still the most crucial factor in the perceptual decision. Therefore, the increased influence of prior choices in ASD does not simply reflect a case of blind repetition. Overall (grouping all stimulus conditions) β_{curr_stimulus} did not differ significantly between ASD and controls (F_1,24 = 0.03, p = 0.88, η_p² = 0.001, Mixed ANOVA). No significant correlation was found between ${\beta }_{prev\_choices}$ and ${\beta }_{curr\_stimulus}$ (r₃₄ = 0.06, p = 0.74). There was no significant correlation between ${\beta }_{prev\_choices}$ and SCQ (r₁₂ = -0.25, p = 0.41 and r₂₀ = 0.12, p = 0.59, for ASD and controls, respectively). Also, no significant correlation was found between ${\beta }_{prev\_choices}$ and age in the ASD group (r₁₂ = 0.54, p = 0.057) or in the control group (r₂₀ = 0.15, p = 0.50).

Effect of recent priors in ASD ensues even when reported using different actions

In order to dissociate the effect of recent choices from motor repetition, an additional location discrimination condition was run – the Response Invariant condition. In this condition, different motor actions were used when responding to ‘prior’ vs. ‘test’ stimuli (identified by different colors, see Materials and methods for more details). Also in this condition ΔPSE values were significantly positive for both the control and ASD groups individually (Figure 7A; t₁₅ = 2.65, p = 0.018, Cohen’s d = 0.66, 95% CI = [0.11, 1.2] and t₁₅ = 2.68, p = 0.017, Cohen’s d = 0.67, 95% CI = [0.12, 1.2], respectively; t-tests). However, there was no significant difference between the groups (Figure 7A; t₃₀ = 0.09, p = 0.93, Cohen’s d = 0.03, 95% CI = [−0.66, 0.73], t-test).

Figure 7 with 2 supplements see all

Download asset Open asset

When comparing the logistic regression coefficients (like above for the primary condition), we found no significant difference between the groups (Figure 7B, F_3,28 = 1.18, p = 0.34, η_p² = 0.11, one-way MANOVA). For both the ASD and control groups, there was a positive effect of previous choices (Figure 7B, center plot; t₁₅ = 3.97, p = 0.001, Cohen’s d = 0.99, 95% CI = [0.38, 1.58] and t₁₅ = 6.73, p = 6.8∙10⁻⁶, Cohen’s d = 1.68, 95% CI = [0.99, 2.44], respectively; t-tests) while the effect of previous stimuli was small and not significant (Figure 7B, right plot; t₁₅ = -1.24, p = 0.23, Cohen’s d = 0.3, 95% CI = [-0.8, 0.2] and t₁₅ = 0.25, p = 0.8, Cohen’s d = 0.06, 95% CI = [-0.43, 0.55], respectively; t-tests). The previous choice coefficients (${\beta }_{prev\_choices}$) did not correlate significantly with ADOS scores (Figure 7—figure supplement 2, r₁₄ = -0.42, p = 0.10). In line with previous findings that using the same action boosts the prior choice effect (Feigin et al., 2021), also here choice biases (${\beta }_{prev\_choices})$ were larger in the primary condition vs. the Response Invariant condition (t₃₀ = 2.94, p = 0.006, Cohen’s d = 0.52, 95% CI = [0.15, 0.89], t-test; BF₀₁ = 0.15; pooling both groups together).

Similar to the primary condition, the current stimulus effect (β_{curr_stimulus}) was large, positive and significant for both groups (Figure 7B, left plot; t₁₅ = 16.26, p = 6.2∙10⁻¹¹, Cohen’s d = 4.07, 95% CI = [2.54, 5.58] and t₁₅ = 10.74, p = 1.9∙10⁻⁸, Cohen’s d = 2.69, 95% CI = [1.6, 3.75], for ASD and controls, respectively; t-tests) and did not differ significantly between the groups (t₃₀ = -1.58, p = 0.125, Cohen’s d = 0.56, 95% CI = [-1.26, 0.15], t-test). This indicates that also in this condition sensitivity did not differ between the groups. We confirmed this by comparing thresholds from the psychometric functions, which also did not differ significantly between groups (t₃₀ = 1.02, p = 0.32, Cohen’s d = 0.36, 95% CI = [-0.34, 1.06], t-test; BF₀₁ = 2.00). Here too, stimulus sensitivity (${\beta }_{curr\_stimulus}$) did not correlate significantly with the prior choice effect ${\beta }_{prev\_choices}$ (r₃₀ = -0.27,p = 0.13), and the baseline coefficients did not differ significantly between the groups (t₃₀ = 1.29, p = 0.2, Cohen’s d = 0.46, 95% CI = [-0.25, 1.15], t-test). Also, response times did not differ between the two groups: mean ± SD = 0.83 ± 0.15s and 0.87 ± 0.12s, for ASD and controls, respectively (calculated across all discriminations, that is, pooling prior and test discriminations; t₃₀ = -0.85, p = 0.4, Cohen’s d = 0.3, 95% CI = [-0.99, 0.40], t-test).

These results show that the positive effect of recent prior choices ensues even when context (stimulus color) and the motor actions used to report responses are altered – also for individuals with ASD. This strengthens the argument against a reduced effect of recent priors in ASD. Taking the results of the two location discrimination conditions (and heading discrimination) together, we conclude that individuals with ASD experience a positive bias from recent choices, like controls. When context is consistent (i.e. stimulus color is the same, and the same motor action is used to report choices) this effect is enlarged in ASD vs. controls. However, when context is changed (and perception-action coupling is broken) this effect is no longer enlarged in ASD (but is nonetheless still present).

In this study, we found that when making perceptual decisions, individuals with ASD are more greatly influenced (vs. controls) by their recent prior choices. For all participants (ASD and control) perceptual decisions were ‘attracted’ toward (i.e. more likely to be the same as) recent prior choices. However, this effect was significantly larger in ASD. This finding was replicated in two different tasks (visual localization and visual/vestibular heading discrimination) performed by two different cohorts – one in Israel and one in the USA. These results are in line with a recent study that showed increased consistency in gambling choices in ASD (Wu et al., 2018). Here, we extend this to perceptual decision-making, and dissociated the effects of prior choices and prior stimulus information.

By contrast, the influence of recent prior sensory information was similar in ASD and controls (in both tasks). This reflected a ‘repulsive’ bias (away from the preceding stimuli) which was large and highly significant for both groups in the heading discrimination task (in the location discrimination task, the effect was similar, but small). Notably, this effect (adaptation to prior stimuli) was in the opposite direction vs. the effect of prior choices. A negative influence of prior stimuli on subsequent perception has been documented in many studies (e.g. sensory adaptation /aftereffects, Gibson, 1937; Gibson and Radner, 1937; Anstis et al., 1998; Thompson, P., and Thompson and Burr, 2009), even with very short (milliseconds) stimulus durations (Sekuler and Littlejohn, 1974; Suzuki, 2001; Felsen et al., 2002; Xu and Liu, 2012; Sou and Xu, 2019). The magnitude of adaptation aftereffects increases with exposure time (Rhodes et al., 2007; Pegors et al., 2015; Burton et al., 2016). This might explain why stimulus adaptation was small in the location discrimination task (0.25 s stimulus duration) and strong (and highly significant) in the heading discrimination task (1 s stimulus duration). Because there was no fixation point for the repeated stimulus presentations in the location discrimination task (all trials in the heading discrimination task had a fixation point), a negative effect of prior stimuli could reflect a shift in the perceived center of the screen.

A positive influence of prior choices has also been documented in many studies (Kaneko and Sakai, 2015; Talluri et al., 2018; Feigin et al., 2021). Furthermore, converging evidence from many recent studies suggests that the influence of prior stimuli and prior choices are often in opposite directions – that is, a negative (adaptive) influence of prior stimuli vs. a positive (consistency bias) influence of prior choices, and that these superimpose (Fritsche et al., 2017; Bosch et al., 2020; Feigin et al., 2021; Sadil et al., 2021). However, opposing effects of prior sensory vs. prior choices on perceptual decisions has not yet been studied in ASD. Importantly, these effects need to be dissociated in order to better understand suggestions of reduced adaptation in ASD (Pellicano et al., 2007; Lawson et al., 2015; Molesworth et al., 2015; Turi et al., 2015; Noel et al., 2017; Lieder et al., 2019).

When separating the effects of prior choices and prior stimuli, we did not find reduced adaptation to prior stimuli in ASD. Our results therefore suggest that observations of ‘reduced adaptation’ in ASD might (at least partially) be explained by an increased positive (i.e. counteracting) effect of prior choices. However, our stimuli were of relatively short duration (0.25 s in the location discrimination task and 1 s in the heading direction discrimination task). Longer stimuli might lead to greater adaptation (Pegors et al., 2015) which might better expose any differences. Thus, future work should test whether adaptation to longer duration stimuli (several seconds) is altered in ASD, while taking into account the effects of prior choices.

The finding of an increased influence of prior choices in ASD may suggest that repetitiveness and inflexibility symptoms might manifest also in perceptual decisions. However, we did not find a significant correlation between the beta coefficients for prior choices and ADOS and ADI-R sub-scores of repetitiveness and inflexibility. Accordingly, our findings might offer an additional measure of repetitive symptoms, which is perhaps not covered by these sub-scores (but this would need a larger data set to confirm). Although there was a trend between the overall effect of priors (ΔPSEs) and the ADI-Communication sub-scores, this did not survive correction for multiple comparisons. Therefore, whether or not an increased influence of prior choices correlates with ASD symptoms warrants further investigation with a larger dataset in the future.

The Response Invariant condition tested the effect of prior choices when context was altered – choices were reported via different motor actions (signaled by different colored stimuli). Also here, individuals with ASD still showed a significant influence of prior choices. This challenges the theories of attenuated priors in ASD (Pellicano and Burr, 2012) and demonstrates that even when context is different, individuals with ASD nonetheless still demonstrate a robust influence of priors (specifically prior choices). However, in this condition, the effect was no longer larger than controls. This could indicate that the primary results reflect ‘motor perseveration’ in ASD rather than a greater influence of prior choices. Although we cannot rule out a contribution of motor perseveration, several additional analyses suggest that it does not fully account for the ASD results. Firstly, a simple motor perseveration effect would manifest irrespective of discrimination difficulty (also for easy discriminations), whereas we found no significant increase in lapse rates in ASD. Also, we found an independent contribution of choices before the previous choice (i.e. two choices back; with significantly larger effects in ASD). This further suggests that the increased effect of prior choices in ASD reflects a compound influence of choices, beyond simple motor repetition.

The observation that an enlarged influence of prior choices in ASD was seen only in the primary condition (and in the heading discrimination condition), when motor responses to prior and test stimuli were the same, but not when the actions for prior and test stimuli were dissociated, suggests that this difference in ASD may be mediated by altered perception-action coupling. Namely, the greater influence of prior choices in ASD may be specific to when perception-action coupling is consistent. This interpretation is in line with the hypothesis that perception and action are intertwined in a perception-action loop (Buckingham et al., 2016) and our recent finding (in a series of related control experiments; Feigin et al., 2021) that sensorimotor (decision-action) coupling boosts the influence of prior choices. Accordingly, the increased influence of prior choices in ASD may reflect an increased influence of prior perception-action coupling on subsequent decisions.

The consistency bias in the heading discrimination task was less strong (and not seen for control participants). We think that this might be because of: (i) relatively long inter-stimulus-intervals (5~6 s, due to technical constraints – the motion platform needs to return slowly to base before a new trial can begin). By contrast, ISIs in the location discrimination task were ~1 s. (ii) The heading discrimination task involved three types of stimuli from two different sensory modalities: visual, vestibular, and combined (both visual and vestibular stimuli delivered together). The stimuli were interleaved, so that in most trials, the current and prior stimulus were of different types. The effect of the previous trial may be decreased when the conditions differ across trials (Feigin et al., 2021).

It has been proposed that individuals with high-functioning ASD have an increased ability and tendency to recognize repeating patterns in stimuli and to generate rules for prediction (Baron-Cohen, 2002; Baron-Cohen, 2009). Our results suggest that an elaborate set of systematic, internal rules in ASD uses information from prior choices to a greater degree. Furthermore, consistency (in context and in the motor responses) might be an important factor for generating prediction rules in ASD, which has been proposed to be a disorder of prediction (Sinha et al., 2014).

In summary, we found here that individuals with ASD demonstrated an increased influence of recent prior choices on perceptual decisions (vs. controls), while the influence of recent stimulus history was unaltered. This was seen in two different tasks using different modalities (visual localization and visual/vestibular heading discrimination), tested in two different cohorts (Israel and USA). These results suggest that perceptual decisions in ASD are more repetitive and less flexible (i.e. more consistent with prior decisions).

We studied serial dependence of perceptual decisions in individuals with ASD vs. controls performing two different tasks: (i) visual location discrimination (designed for this study and performed in Israel), (ii) multisensory (visual-vestibular) heading discrimination (reanalysis of previously published data, collected in the USA; Zaidel et al., 2015). We present below participant and task details of the former (visual location discrimination), and briefly summarize relevant details of the latter (heading discrimination), for which full details are available in the original publication.

The data and analysis code for the location discrimination experiments (i.e., the new data from this study) have been uploaded to github and can be found at https://github.com/HF-GH/ASD (copy archived at https://archive.softwareheritage.org/swh:1:rev:313acbcb7ec08606d1a3332c6ee8645a091aae1a).

1. 1. Alexi J
   2. Cleary D
   3. Dommisse K
   4. Palermo R
   5. Kloth N
   6. Burr D
   7. Bell J

   (2018) Past visual experiences weigh in on body size estimation

   Scientific Reports 8:1–8.

   https://doi.org/10.1038/s41598-017-18418-3

   • PubMed
   • Google Scholar
2. Book

   1. American Psychiatric Association

   (2000)

   Diagnostic and Statistical Manual of Mental Disorders (4th ed)

   Washington, DC: American Psychiatric Publishing.

   • Google Scholar
3. Book

   1. American Psychiatric Association

   (2013)

   The Diagnostic and Statistical Manual of Mental Disorders: DSM-5 (5th ed)

   Washington, DC: American Psychiatric Publishing.

   • Google Scholar
4. 1. Anstis S
   2. Verstraten FA
   3. Mather G

   (1998) The motion aftereffect

   Trends in Cognitive Sciences 2:111–117.

   https://doi.org/10.1016/s1364-6613(98)01142-5

   • PubMed
   • Google Scholar
5. 1. Baron-Cohen S

   (2002) The extreme male brain theory of autism

   Trends in Cognitive Sciences 6:248–254.

   https://doi.org/10.1016/s1364-6613(02)01904-6

   • PubMed
   • Google Scholar
6. 1. Baron-Cohen S

   (2009) Autism: the empathizing-systemizing (E-S) theory

   Annals of the New York Academy of Sciences 1156:68–80.

   https://doi.org/10.1111/j.1749-6632.2009.04467.x

   • PubMed
   • Google Scholar
7. 1. Behrmann M
   2. Avidan G
   3. Leonard GL
   4. Kimchi R
   5. Luna B
   6. Humphreys K
   7. Minshew N

   (2006) Configural processing in autism and its relationship to face processing

   Neuropsychologia 44:110–129.

   https://doi.org/10.1016/j.neuropsychologia.2005.04.002

   • PubMed
   • Google Scholar
8. 1. Booth RDL
   2. Happé FGE

   (2018) Evidence of reduced global processing in autism spectrum disorder

   Journal of Autism and Developmental Disorders 48:1397–1408.

   https://doi.org/10.1007/s10803-016-2724-6

   • PubMed
   • Google Scholar
9. 1. Bosch E
   2. Fritsche M
   3. Ehinger BV
   4. de Lange FP

   (2020) Opposite effects of choice history and evidence history resolve a paradox of sequential choice Bias

   Journal of Vision 20:9–13.

   https://doi.org/10.1167/jov.20.12.9

   • Google Scholar
10. 1. Brock J

    (2012) Alternative Bayesian accounts of autistic perception: comment on Pellicano and Burr

    Trends in Cognitive Sciences 16:573–574.

    https://doi.org/10.1016/j.tics.2012.10.005

    • PubMed
    • Google Scholar
11. 1. Buckingham G
    2. Michelakakis EE
    3. Rajendran G

    (2016) The influence of prior knowledge on perception and action: relationships to autistic traits

    Journal of Autism and Developmental Disorders 46:1716–1724.

    https://doi.org/10.1007/s10803-016-2701-0

    • PubMed
    • Google Scholar
12. 1. Burton N
    2. Jeffery L
    3. Bonner J
    4. Rhodes G

    (2016) The timecourse of expression aftereffects

    Journal of Vision 16:1.

    https://doi.org/10.1167/16.15.1

    • PubMed
    • Google Scholar
13. 1. Cornsweet TN

    (1962) The staircrase-method in psychophysics

    The American Journal of Psychology 75:485–491.

    https://doi.org/10.2307/1419876

    • PubMed
    • Google Scholar
14. 1. Croydon A
    2. Karaminis T
    3. Neil L
    4. Burr D
    5. Pellicano E

    (2017) The light-from-above prior is intact in autistic children

    Journal of Experimental Child Psychology 161:113–125.

    https://doi.org/10.1016/j.jecp.2017.04.005

    • PubMed
    • Google Scholar
15. Software

    1. Feigin H

    (2021) ASD - location discrimination task, version swh:1:rev:313acbcb7ec08606d1a3332c6ee8645a091aae1a

    Software Heritage.

    https://archive.softwareheritage.org/swh:1:rev:313acbcb7ec08606d1a3332c6ee8645a091aae1a
16. 1. Feigin H
    2. Baror S
    3. Bar M
    4. Zaidel A

    (2021) Perceptual decisions are biased toward relevant prior choices

    Scientific Reports 11:1–16.

    https://doi.org/10.1038/s41598-020-80128-0

    • PubMed
    • Google Scholar
17. 1. Felsen G
    2. Shen YS
    3. Yao H
    4. Spor G
    5. Li C
    6. Dan Y

    (2002) Dynamic modification of cortical orientation tuning mediated by recurrent connections

    Neuron 36:945–954.

    https://doi.org/10.1016/s0896-6273(02)01011-5

    • PubMed
    • Google Scholar
18. 1. Fischer J
    2. Whitney D

    (2014) Serial dependence in visual perception

    Nature Neuroscience 17:738–743.

    https://doi.org/10.1038/nn.3689

    • PubMed
    • Google Scholar
19. 1. Friston KJ
    2. Lawson R
    3. Frith CD

    (2013) On hyperpriors and hypopriors: comment on Pellicano and burr

    Trends in Cognitive Sciences 17:1.

    https://doi.org/10.1016/j.tics.2012.11.003

    • PubMed
    • Google Scholar
20. 1. Frith U
    2. Happé F

    (1994) Autism: beyond “theory of mind”

    Cognition 50:115–132.

    https://doi.org/10.1016/0010-0277(94)90024-8

    • Google Scholar
21. 1. Fritsche M
    2. Mostert P
    3. de Lange FP

    (2017) Opposite effects of recent history on perception and decision

    Current Biology 27:590–595.

    https://doi.org/10.1016/j.cub.2017.01.006

    • PubMed
    • Google Scholar
22. 1. Gibson JJ

    (1937) Adaptation with negative after-effect

    Psychological Review 44:222–244.

    https://doi.org/10.1037/h0061358

    • Google Scholar
23. 1. Gibson JJ
    2. Radner M

    (1937) Adaptation, after-effect and contrast in the perception of tilted lines

    Journal of Experimental Psychology 20:453–467.

    https://doi.org/10.1037/h0059826

    • Google Scholar
24. 1. Hadad BS
    2. Schwartz S

    (2019) Perception in autism does not adhere to Weber's law

    eLife 8:e42223.

    https://doi.org/10.7554/eLife.42223

    • PubMed
    • Google Scholar
25. 1. Happé F
    2. Frith U

    (2006) The weak coherence account: detail-focused cognitive style in autism spectrum disorders

    Journal of Autism and Developmental Disorders 36:5–25.

    https://doi.org/10.1007/s10803-005-0039-0

    • PubMed
    • Google Scholar
26. Book

    1. Hosmer DW
    2. Lemeshow S

    (1989) Applied Logistic Regression

    New York: Wiley.

    https://doi.org/10.1002/9781118548387

    • Google Scholar
27. 1. Kaneko Y
    2. Sakai K

    (2015) Dissociation in decision Bias mechanism between probabilistic information and previous decision

    Frontiers in Human Neuroscience 9:1–11.

    https://doi.org/10.3389/fnhum.2015.00261

    • PubMed
    • Google Scholar
28. 1. Karvelis P
    2. Seitz AR
    3. Lawrie SM
    4. Seriès P

    (2018) Autistic traits, but not schizotypy, predict increased weighting of sensory information in bayesian visual integration

    eLife 7:e34115.

    https://doi.org/10.7554/eLife.34115

    • Google Scholar
29. 1. Kersten D
    2. Yuille A

    (2003) Bayesian models of object perception

    Current Opinion in Neurobiology 13:150–158.

    https://doi.org/10.1016/S0959-4388(03)00042-4

    • PubMed
    • Google Scholar
30. 1. Knill DC
    2. Pouget A

    (2004) The bayesian brain: the role of uncertainty in neural coding and computation

    Trends in Neurosciences 27:712–719.

    https://doi.org/10.1016/j.tins.2004.10.007

    • PubMed
    • Google Scholar
31. Book

    1. Knill D
    2. Richards W

    (1996)

    Perception as Bayesian Inference

    Cambridge University Press.

    • Google Scholar
32. 1. Kohn A

    (2007) Visual adaptation: physiology, mechanisms, and functional benefits

    Journal of Neurophysiology 97:3155–3164.

    https://doi.org/10.1152/jn.00086.2007

    • Google Scholar
33. 1. Koldewyn K
    2. Jiang YV
    3. Weigelt S
    4. Kanwisher N

    (2013) Global/local processing in autism: not a disability, but a disinclination

    Journal of Autism and Developmental Disorders 43:2329–2340.

    https://doi.org/10.1007/s10803-013-1777-z

    • PubMed
    • Google Scholar
34. 1. Lawson RP
    2. Rees G
    3. Friston KJ

    (2014) An aberrant precision account of autism

    Frontiers in Human Neuroscience 8:302.

    https://doi.org/10.3389/fnhum.2014.00302

    • PubMed
    • Google Scholar
35. 1. Lawson RP
    2. Aylward J
    3. White S
    4. Rees G

    (2015) A striking reduction of simple loudness adaptation in autism

    Scientific Reports 5:1–7.

    https://doi.org/10.1038/srep16157

    • PubMed
    • Google Scholar
36. 1. Leek MR

    (2001) Adaptive procedures in psychophysical research

    Perception & Psychophysics 63:1279–1292.

    https://doi.org/10.3758/bf03194543

    • PubMed
    • Google Scholar
37. 1. Liberman A
    2. Fischer J
    3. Whitney D

    (2014) Serial dependence in the perception of faces

    Current Biology: CB 24:2569–2574.

    https://doi.org/10.1016/j.cub.2014.09.025

    • PubMed
    • Google Scholar
38. 1. Liberman A
    2. Manassi M
    3. Whitney D

    (2018) Serial dependence promotes the stability of perceived emotional expression depending on face similarity

    Attention, Perception, & Psychophysics 80:1461–1473.

    https://doi.org/10.3758/s13414-018-1533-8

    • Google Scholar
39. 1. Lieder I
    2. Adam V
    3. Frenkel O
    4. Jaffe-Dax S
    5. Sahani M
    6. Ahissar M

    (2019) Perceptual Bias reveals slow-updating in autism and fast-forgetting in dyslexia

    Nature Neuroscience 22:256–264.

    https://doi.org/10.1038/s41593-018-0308-9

    • PubMed
    • Google Scholar
40. 1. Lord C
    2. Rutter M
    3. Le Couteur A

    (1994) Autism diagnostic Interview-Revised: a revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders

    Journal of Autism and Developmental Disorders 24:659–685.

    https://doi.org/10.1007/BF02172145

    • Google Scholar
41. Book

    1. Lord C
    2. Rutter M
    3. DiLavore PC
    4. Risi S
    5. Gotham K
    6. Bishop S

    (2012)

    Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) Manual (Part I): Modules 1–4

    Torrance, CA: Western Psychological Services.

    • Google Scholar
42. 1. Ma WJ

    (2019) Bayesian decision models: A primer

    Neuron 104:164–175.

    https://doi.org/10.1016/j.neuron.2019.09.037

    • PubMed
    • Google Scholar
43. Book

    1. Mamassian P
    2. Landy M
    3. Maloney LT

    (2018)

    Bayesian Modelling of Visual Perception

    In: Mamassian P, editors. Probabilistic Models of the Brain. MIT Press. pp. 13–36.

    • Google Scholar
44. 1. Menard S

    (2000) Coefficients of determination for multiple logistic regression analysis

    The American Statistician 54:17–24.

    https://doi.org/10.1080/00031305.2000.10474502

    • Google Scholar
45. 1. Molesworth C
    2. Chevallier C
    3. Happé F
    4. Hampton JA

    (2015) Children with autism do not show sequence effects with auditory stimuli

    Journal of Experimental Psychology: General 144:48–57.

    https://doi.org/10.1037/a0038204

    • PubMed
    • Google Scholar
46. 1. Mottron L
    2. Dawson M
    3. Soulières I
    4. Hubert B
    5. Burack J

    (2006) Enhanced perceptual functioning in autism: an update, and eight principles of autistic perception

    Journal of Autism and Developmental Disorders 36:27–43.

    https://doi.org/10.1007/s10803-005-0040-7

    • PubMed
    • Google Scholar
47. Book

    1. Mottron L
    2. Burack J

    (2021) Enhanced Perceptual Functioning in the Development of Autism

    In: Mottron L, editors. The Development of Autism. Springer. pp. 136–151.

    https://doi.org/10.1007/978-1-4419-1698-3_723

    • Google Scholar
48. 1. Noel JP
    2. De Niear MA
    3. Stevenson R
    4. Alais D
    5. Wallace MT

    (2017) Atypical rapid audio-visual temporal recalibration in autism spectrum disorders

    Autism Research 10:121–129.

    https://doi.org/10.1002/aur.1633

    • PubMed
    • Google Scholar
49. 1. Noel JP
    2. Lakshminarasimhan KJ
    3. Park H
    4. Angelaki DE

    (2020) Increased variability but intact integration during visual navigation in autism spectrum disorder

    PNAS 117:11158–11166.

    https://doi.org/10.1073/pnas.2000216117

    • PubMed
    • Google Scholar
50. 1. Pascucci D
    2. Mancuso G
    3. Santandrea E
    4. Della Libera C
    5. Plomp G
    6. Chelazzi L

    (2019) Laws of concatenated perception: Vision goes for novelty, decisions for perseverance

    PLOS biology 17:e3000144.

    https://doi.org/10.1371/journal.pbio.3000144

    • PubMed
    • Google Scholar
51. 1. Pegors TK
    2. Mattar MG
    3. Bryan PB
    4. Epstein RA

    (2015) Simultaneous perceptual and response biases on sequential face attractiveness judgments

    Journal of Experimental Psychology: General 144:664–673.

    https://doi.org/10.1037/xge0000069

    • PubMed
    • Google Scholar
52. 1. Peirce JW

    (2007) PsychoPy--psychophysics software in python

    Journal of Neuroscience Methods 162:8–13.

    https://doi.org/10.1016/j.jneumeth.2006.11.017

    • PubMed
    • Google Scholar
53. 1. Pellicano E
    2. Jeffery L
    3. Burr D
    4. Rhodes G

    (2007) Abnormal adaptive Face-Coding mechanisms in children with autism spectrum disorder

    Current Biology 17:1508–1512.

    https://doi.org/10.1016/j.cub.2007.07.065

    • Google Scholar
54. 1. Pellicano E
    2. Burr D

    (2012) When the world becomes 'too real': a Bayesian explanation of autistic perception

    Trends in Cognitive Sciences 16:504–510.

    https://doi.org/10.1016/j.tics.2012.08.009

    • PubMed
    • Google Scholar
55. 1. Peters MA
    2. Ma WJ
    3. Shams L

    (2016) The Size-Weight illusion is not anti-Bayesian after all: a unifying bayesian account

    PeerJ 4:e2124.

    https://doi.org/10.7717/peerj.2124

    • PubMed
    • Google Scholar
56. 1. Plaisted K
    2. Swettenham J
    3. Rees L

    (1999) Children with autism show local precedence in a divided attention task and global precedence in a selective attention task

    Journal of Child Psychology and Psychiatry 40:733–742.

    https://doi.org/10.1111/1469-7610.00489

    • PubMed
    • Google Scholar
57. 1. Plaisted K
    2. Saksida L
    3. Alcántara J
    4. Weisblatt E

    (2003) Towards an understanding of the mechanisms of weak central coherence effects: experiments in visual configural learning and auditory perception

    Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 358:375–386.

    https://doi.org/10.1098/rstb.2002.1211

    • Google Scholar
58. 1. Rhodes G
    2. Jeffery L
    3. Clifford CW
    4. Leopold DA

    (2007) The timecourse of higher-level face aftereffects

    Vision Research 47:2291–2296.

    https://doi.org/10.1016/j.visres.2007.05.012

    • PubMed
    • Google Scholar
59. 1. Rogers SJ
    2. Ozonoff S

    (2005) Annotation: what do we know about sensory dysfunction in autism? A critical review of the empirical evidence

    Journal of Child Psychology and Psychiatry 46:1255–1268.

    https://doi.org/10.1111/j.1469-7610.2005.01431.x

    • PubMed
    • Google Scholar
60. Book

    1. Rutter M
    2. Bailey A
    3. Lord C

    (2003a)

    The Social Communication Questionnaire: Manual

    Los Angeles, CA: Western Psychological Services.

    • Google Scholar
61. Book

    1. Rutter M
    2. Le Couteur A
    3. Lord C

    (2003b)

    Autism Diagnostic Interview-Revised

    Los Angeles, CA: Western Psychological Services.

    • Google Scholar
62. Preprint

    1. Sadil P
    2. Cowell R
    3. Huber DE

    (2021) The Yin-yang of serial dependence effects: every response is both an attraction to the prior response and a repulsion from the prior stimulus

    PsyArXiv.

    https://psyarxiv.com/f52yz/

    • Google Scholar
63. 1. Schütt HH
    2. Harmeling S
    3. Macke JH
    4. Wichmann FA

    (2016) Painfree and accurate bayesian estimation of psychometric functions for (potentially) overdispersed data

    Vision Research 122:105–123.

    https://doi.org/10.1016/j.visres.2016.02.002

    • PubMed
    • Google Scholar
64. 1. Sekuler R
    2. Littlejohn J

    (1974) Letter: tilt aftereffect following very brief exposures

    Vision Research 14:151–152.

    https://doi.org/10.1016/0042-6989(74)90133-3

    • PubMed
    • Google Scholar
65. 1. Shah A
    2. Frith U

    (1983) An islet of ability in autistic children: a research note

    Journal of Child Psychology and Psychiatry 24:613–620.

    https://doi.org/10.1111/j.1469-7610.1983.tb00137.x

    • PubMed
    • Google Scholar
66. 1. Sinha P
    2. Kjelgaard MM
    3. Gandhi TK
    4. Tsourides K
    5. Cardinaux AL
    6. Pantazis D
    7. Diamond SP
    8. Held RM

    (2014) Autism as a disorder of prediction

    PNAS 111:15220–15225.

    https://doi.org/10.1073/pnas.1416797111

    • PubMed
    • Google Scholar
67. 1. Sou KL
    2. Xu H

    (2019) Brief facial emotion aftereffect occurs earlier for angry than happy adaptation

    Vision Research 162:35–42.

    https://doi.org/10.1016/j.visres.2019.07.002

    • PubMed
    • Google Scholar
68. 1. Suzuki S

    (2001) Attention-dependent brief adaptation to contour orientation: a high-level aftereffect for convexity?

    Vision Research 41:3883–3902.

    https://doi.org/10.1016/S0042-6989(01)00249-8

    • PubMed
    • Google Scholar
69. 1. Talluri BC
    2. Urai AE
    3. Tsetsos K
    4. Usher M
    5. Donner TH

    (2018) Confirmation Bias through selective overweighting of Choice-Consistent evidence

    Current Biology 28:3128–3135.

    https://doi.org/10.1016/j.cub.2018.07.052

    • PubMed
    • Google Scholar
70. 1. Taubert J
    2. Van der Burg E
    3. Alais D

    (2016) Love at second sight: sequential dependence of facial attractiveness in an on-line dating paradigm

    Scientific Reports 6:2–6.

    https://doi.org/10.1038/srep22740

    • PubMed
    • Google Scholar
71. 1. Thompson P
    2. Burr D

    (2009) Visual aftereffects

    Current Biology 19:R11–R14.

    https://doi.org/10.1016/j.cub.2008.10.014

    • PubMed
    • Google Scholar
72. 1. Turi M
    2. Burr DC
    3. Igliozzi R
    4. Aagten-Murphy D
    5. Muratori F
    6. Pellicano E

    (2015) Children with autism spectrum disorder show reduced adaptation to number

    PNAS 112:7868–7872.

    https://doi.org/10.1073/pnas.1504099112

    • PubMed
    • Google Scholar
73. 1. Van de Cruys S
    2. Vanmarcke S
    3. Van de Put I
    4. Wagemans J

    (2018) The use of prior knowledge for perceptual inference is preserved in ASD

    Clinical Psychological Science 6:382–393.

    https://doi.org/10.1177/2167702617740955

    • Google Scholar
74. Book

    1. Wechsler D

    (2010)

    WISC-IV HEB. Wechsler Intelligence Scale for Children

    Jerusalem, Israel: Psychtech.

    • Google Scholar
75. 1. Wu HC
    2. White S
    3. Rees G
    4. Burgess PW

    (2018) Executive function in high-functioning autism: decision-making consistency as a characteristic gambling behaviour

    Cortex; a Journal Devoted to the Study of the Nervous System and Behavior 107:21–36.

    https://doi.org/10.1016/j.cortex.2018.01.013

    • PubMed
    • Google Scholar
76. 1. Xu H
    2. Liu P

    (2012) Adapting to an incomplete curve generates the same curvature aftereffect as a complete curve

    Journal of Vision 12:1052.

    https://doi.org/10.1167/12.9.1052

    • Google Scholar
77. 1. Zaidel A
    2. Goin-Kochel RP
    3. Angelaki DE

    (2015) Self-motion perception in autism is compromised by visual noise but integrated optimally across multiple senses

    PNAS 112:6461–6466.

    https://doi.org/10.1073/pnas.1506582112

    • PubMed
    • Google Scholar

Author details

1. Helen Feigin

   Gonda Multidisciplinary Brain Research Center, Bar Ilan University, Ramat Gan, Israel

   Contribution

   Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft

   Contributed equally with

   Shir Shalom-Sperber

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-6069-0416

2. Shir Shalom-Sperber

   Gonda Multidisciplinary Brain Research Center, Bar Ilan University, Ramat Gan, Israel

   Contribution

   Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft

   Contributed equally with

   Helen Feigin

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-2455-5871

3. Ditza A Zachor

   1. The Autism Center/ALUT, Shamir Medical Center, Zerifin, Israel
   2. Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel

   Contribution

   Resources, Data curation, Writing - review and editing

   Competing interests

   No competing interests declared

4. Adam Zaidel

   Gonda Multidisciplinary Brain Research Center, Bar Ilan University, Ramat Gan, Israel

   Contribution

   Conceptualization, Resources, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing - original draft, Project administration, Writing - review and editing

   For correspondence

   adam.zaidel@biu.ac.il

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4405-8717

• 1,987

  views

• 294

  downloads

• 15

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.