Comparative Study of Predictive Task Performance in ASD vs. NT Children with AI Enhancement
Poster No:
331
Submission Type:
Abstract Submission
Authors:
Chayan Majumder1, Amit Bhongade1, Sheffali Gulati2, Tapan Gandhi1
Institutions:
1Indian Institute of Technology Delhi, South West Delhi, DELHI, 2AIIMS Delhi, South West Delhi, DELHI
First Author:
Co-Author(s):
Introduction:
Autism Spectrum Disorder (ASD) is a neurodevelopmental condition marked by diverse behavioral and functional challenges, varying widely among individuals. A key hypothesis links ASD to impairments in the brain's predictive abilities-anticipating sensory stimuli, events, and social cues [1][2]. Predictive processing is essential for environmental navigation and social interaction, yet it remains unclear how deficits in spatial (location of objects) and temporal (timing of events) prediction contribute to ASD. While both may be affected, their specific roles are uncertain. This study explores whether basic temporal and spatial prediction abilities are impaired in ASD using machine-learning-based interactive video games and whether AI tools can improve these abilities.
Methods:
The study involved 12 children with ASD and 12 age- and IQ-matched neurotypical (NT) controls, aged 4–7 years, with CARS scores in the mild-to-moderate range. Interactive video games based on a reinforcement learning framework were developed to assess temporal (e.g., duration of screen touch) and spatial (e.g., stick length estimation) prediction abilities. Game parameters were dynamically updated using machine learning techniques based on participants' prior responses. The games were presented on 8-inch Android-based Lenovo tablets with sound effects to enhance engagement. A 20-minute priming session, supported by parents and psychologists, was conducted to familiarize participants with the task. Following this, participants completed the tasks while seated comfortably in a distraction-free environment. Performance data were recorded on remote servers, and tasks were dynamically adapted to improve prediction abilities. The data processing workflow is shown in Figure 1.
·Complete data processing block diagram for prediction ability estimation in ASD and NT subjects
Results:
Significant differences were observed between ASD and NT subjects aged 4–7 years across key performance measures, including trials completed, prediction accuracy, and task duration, as shown in Figure 2. ASD subjects completed fewer trials (18.6 per domain) compared to NT subjects (31.1) within an 80-second play block and exhibited shorter attention spans. Task duration was also significantly longer for ASD subjects, averaging 30.66 ± 5.58 minutes, versus 17 ± 2 minutes for NT subjects. Additionally, ASD participants completed an average of 2.9 consecutive challenges without breaks exceeding 4 seconds, compared to 5.6 for NT subjects. Prediction accuracy was notably lower in ASD subjects, with an average success rate of 22.29 ± 7.52%, significantly lower than 63.84 ± 6.27% for NT subjects (p < 0.01). While no significant correlation was found between performance outcomes and CARS scores, ASD subjects with better eye-contact scores performed comparatively better. Furthermore, affective states, particularly connected to parental bonds, appeared to influence predictive abilities in ASD participants. The AI-based reinforcement learning model demonstrated its efficacy, leading to a 21% improvement in spatial and temporal prediction performance across successive trials.
·The number of tries and correct prediction accuracy for ASD and NT subjects.
Conclusions:
The study found significant differences between ASD and NT children (aged 4–7 years) in task engagement, prediction accuracy, and cognitive performance. ASD participants had shorter attention spans, fewer trials, and lower prediction accuracy, suggesting challenges in spatial and temporal prediction. Performance did not correlate with CARS scores, but better eye-contact scores were linked to improved predictive performance. Emotional states, especially familial bonds, influenced cognitive tasks in ASD children. The AI-based reinforcement learning model enhanced predictive abilities by 21%. These findings highlight the potential of reinforcement-based interventions to improve cognitive processing and attention in ASD children. Future research should explore long-term effects and underlying mechanisms.
Disorders of the Nervous System:
Neurodevelopmental/ Early Life (eg. ADHD, autism) 1
Perception, Attention and Motor Behavior:
Attention: Visual 2
Keywords:
Autism
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
2. Gandhi, T. K., Tsourides, K., Singhal, N., Cardinaux, A., Jamal, W., Pantazis, D., ... & Sinha, P. (2021). Autonomic and electrophysiological evidence for reduced auditory habituation in autism. Journal of Autism and Developmental Disorders, 51, 2218-2228.