Mechanistic computational models of cognition across scales and species.
TL;DR
The promise of computational approaches via three recent trends are highlighted: multi-scale models of working memory which link receptors and cell-types to brain-wide activity and behavior; the extension of mechanistic computational models from working memory to other cognitive functions including conscious perception, attention and planning; and cross-species comparative models of cognition that give rise to translational computational neuroscience.
Abstract
How can theory and mathematical modeling help accelerate research in psychiatry? Here we highlight the promise of computational approaches via three recent trends: multi-scale models of working memory which link receptors and cell-types to brain-wide activity and behavior; the extension of mechanistic computational models from working memory to other cognitive functions including conscious perception, attention and planning; and cross-species comparative models of cognition that give rise to translational computational neuroscience. These developments demonstrate the feasibility of achieving understanding across spatial and biological scales in psychiatry research. Crucially, computational findings now can give rise to specific and testable mechanistic hypotheses for a wider range of symptoms associated with mental disorders. Finally, computational tools have emerged to guide translational neuroscience towards the experiments that are most likely to translate to treatment of human brain disease. We provide primer coding notebooks for several models to accompany the article.