By April Cashin-Garbutt

Autism spectrum disorder (ASD) is often manifested by social difficulties. Social situations are very dynamic, which means that our predictions and plans need to be constantly updated as we respond to online feedback. Take having a conversation for example, sometimes someone will interject and the topic of discussion will change direction.

“During a conversation, you’re constantly having to update your world model of what is going on. Even at a perceptual level, autistic participants seem to be slower at updating such world models,” explained Dr Nate Zuk, Senior Lecturer at Nottingham Trent University and previously an ELSC-SWC postdoctoral fellow at Hebrew University co-supervised by Professor Merav Ahissar and SWC Group Leader Dr Athena Akrami.

To understand how the brain responds to feedback, Dr Zuk set up a perceptual task for participants and used electroencephalography (EEG) to measure their brain activity related to an area called the anterior cingulate cortex (ACC), which is known to be crucial in dynamic environments, and is also known to produce a feedback related response. Participants included both neurotypical individuals and people with autism. The study aimed to compare the neural responses to feedback on perceptual performance of these two groups. These responses provide insights into differences in updating perceptual priors and testing whether reduced error monitoring could act as a reliable biomarker for autism.

Two-tone perceptual task

Participants in each trial were presented with two sequential auditory tones, separated by a working memory delay interval, and they had to decide whether the first or the second was higher in pitch. The team studied whether participants’ decisions were affected by prior information, particularly when the tones were very close together in pitch resulting in a difficult decision.

This work builds upon on previous research in Ahissar’s and Akrami’s labs. Previous research from the Akrami lab uncovered how central tendency bias, the phenomenon where people’s judgements are biased towards the average of previous observations, is related to recency bias in working memory. Earlier work from the Ahissar lab showed that in neurotypical people, recent events are given particularly larger weight in forming priors, allowing participants to update more quickly in dynamic environments. In autism, the effect of recent events is reduced.

“In our 2019 study, in collaboration with GCNU, we found that more recent events are underweighted in autistic individuals compared with the general population, slowing updating in dynamic environments . In this study, we wanted to see how responses to feedback are produced and how this compares between autistic and non-autistic participants,” explained Professor Ahissar, Hebrew University and co-author of the study.

Responding to feedback

Feedback is essential for adapting to new situations as it helps us to adjust our behaviour accordingly. In this study, the researchers looked at both external feedback (whether the trial was correct or incorrect, indicated by the experimenters) and internal feedback (whether the trial was considered easy or difficult).

“When we know a task is easy and we get it correct, it gives us higher confidence and allows us weigh prior beliefs and expectations as stronger. But in the autistic population, this bias is weaker following an easy and correct trial. This correlates with a lack of feedback-related negativity signal,” explained Dr Akrami.

Feedback-related negativity (FRN)

By using ACC-related EEG signals to measure the response to feedback in perceptual tasks, the researchers showed ASD participants had a reduced response to external feedback. This EEG signal following “incorrect” feedback is known as feedback-related negativity (FRN) and was found to be weaker for autistic individuals, in fast motor tasks. Importantly, FRN has never been assessed, neither in autistics nor non-autistics, in perceptual tasks, where performance following a perceptual success or error can be modified by learning, updating and incorporating internal models.

“For the first time, we measured FRN in a perceptual decision making paradigm, and we were particularly surprised by how consistent this error-related response was across the participants. Previous studies using easy, non-perceptual tasks had found mixed results leading to the hypothesis that responses to, for instance, external feedback of monetary rewards are typical in autism. Therefore, we weren’t expecting the group difference to be so striking in our task. But our results showed that we can reliably measure this reduced tracking of external feedback” explained Dr Zuk.

“We believe that the crucial point is what participants do with the feedback. Our current hypothesis is that in dynamic environments where feedback can be used to change behaviour, there will be significant differences between autistics and neurotypicals, both in FRN and response to feedback”, continued Dr Akrami.

Clinical implications

The findings of this study have important implications for understanding ASD. The weaker processing of both external and internal feedback in autistic participants may contribute to the challenges they face in social communication and interactions. This opens new avenues for study.

In addition, the FRN signal could potentially be used in a clinical setting as a neuropsychological measure of error processing. The Ahissar lab are now doing follow-up studies to discover whether any manipulations to the type and characteristics of feedback can make it more effective for people with autism.

“There is some indirect evidence that the reward system is typical in the autistic population but does not necessarily take part in perceptual tasks, and so we are now testing whether the reward system could be harnessed to make performance monitoring more efficient. This could potentially have an amazing impact on the lives of autistic people,” concluded Dr Ahissar.

Find out more

• Read the full preprint in bioRxiv, ‘Reduced monitoring of task performance is an effective biomarker of autism’ DOI: 10.1101/2024.07.23.604831
• Learn more about research by Nathaniel Zuk, the Ahissar Lab and Akrami Lab

Banner image by Teeradej / Shutterstock.com