Introduction Drift Diffusion (DD) models and spiking neural networks have been effective in predicting and simulating disordered behaviour. They offer deep insight into behavioural processes and spiking neuronal models understand the relationship between such processes and the underlying biological system. A joint approach is valuable as DD models reflect the entirety of performance data and spiking networks mimic neuronal communication. This offers a more detailed analysis that is biologically-plausible and gives a deeper insight into neural dynamics and cognitive processes. Thus, our work uses both models to identify differences among fetal alcohol spectrum disorders (FASD), attention-deficit/hyperactivity disorder (ADHD), and comorbid presentations.
Methods DD models assume, following encoding, that decisions are made via a noisy process where evidence for a response is accumulated until it crosses a response boundary, after which a corresponding motor response is initiated. DD analysis is used to identify how attention processes differ and to inform spiking Search over Time and Space (sSoTS) model parameters. sSoTS is a spiking neuronal model including several synaptic components (AMPA, NMDA, GABA) and frequency adaptation mechanisms. Pool coupling parameter is changed to simulate differences in encoding between FASD and ADHD. Furthermore, we extend prior research in visual search [1,2] by simulating visual selective attention processes to distinguish FASD, ADHD, and comorbid presentations.
Results DD analysis showed ADHD favoured accuracy, whereas FASD (without ADHD) had a similar boundary separation to controls on easy search. ADHD and controls had a similar drift rate, whereas FASD (with and without ADHD) had a slower drift rate on easy visual search. On difficult search, all participants had a similar boundary separation favouring accuracy. All diagnostic groups had similar drift rates on difficult search, which was slower than controls. Our results suggest that FASD affect bottom-up attention processes but an ADHD comorbidity may buffer some effects. To further understand how attention processes differ, data from DD analysis was used to inform sSoTS parameters and changes were successfully simulated.
Discussion Results from our visual search paradigm demonstrated the importance of combining computational models to distinguish different patterns of attention deficit among individuals with FASD, ADHD, and comorbid presentations. Incorporating coupling parameter changes into the sSoTS model successfully simulated the observed behaviours. Our outcomes provide an initial demonstration of how integrating computational methodologies can further enhance our understanding of how attention processes differ across different disorders. Our work underscores that FASD, ADHD, and when both disorders present comorbidly may be able to be distinguished based on the efficiency of bottom-up attention processes.
References [1] Mavritsaki, E., Heinke, D., Allen, H., Deco, G., & Humphreys, G. W. (2011). Bridging the Gap Between Physiology and Behavior: Evidence from the sSoTS Model of Human Visual Attention. Psychological Review, 118(1), 3–41. https://doi.org/10.1037/a0021868 [2] Mavritsaki, E., & Humphreys, G. (2016). Temporal Binding and Segmentation in Visual Search: A Computational Neuroscience Analysis. Journal of Cognitive Neuroscience, 28(10), 1553–1567. https://doi.org/10.1162/jocn_a_00984
Acknowledgement I would like to thank my supervisory team and all collaborators for their support in preparing this abstract.
