Hosted by the Wellcome Neural Dynamics PhD Programme
Studies of human cognition commonly assume our operationalizations must be simple if our inferences are to be intelligible. But not all parts of the description of a complex ability are plausibly intuitively reducible, and insisting on it likely places an artificially low ceiling on the fidelity of our models of the brain. Focusing on tests of fluency and fluid intelligence, here I show how deep generative models may cast light on the nature of neuropsychological tests, enhance their anatomical localising power, and clarify their relation to cognitive abilities they seek to measure.
Amy Nelson is a Doctor and Data Scientist at UCL Department of Brain Repair & Rehabilitation. Her research interests include optimising hospital systems using machine learning, and developing novel metrics of research impact based on predictive modelling.
If joining online: https://bristol-ac-uk.zoom.us/j/94138286231?pwd=MlRURE1SWjR6OTZCR1Fnak9QbGxhUT09, Passcode: 277162