To stay the same the brain has to keep changing. This paradox could explain why functional alterations at the neuronal level, often accompanied by a deviation from a typical behavior, remain of no practical importance for the clinician. Most neurological symptoms are still assessed –and treated– at the behavioral level, whereas their causes, typically lying well within the nervous system, are not directly accessible. Since deeper understanding would allow the development of improved clinical solutions, our clinical data-driven models and algorithms aim to reveal the neurophysiological underpinnings of behavioral impairments.

In this talk, I will present the most recent parts of this work: A Hodgkin-Huxley type model of the sensorimotor cortex that receives neural spikes as inputs and outputs an MEG signal, is used to explain the differences in the MEG recordings between an autistic and a typically developing brain; a distance-wise weighing of the pathophysiological beta band peaks from the microelectrode recordings of the subthalamic nucleus predicts, intraoperatively, the long-term effect of deep brain stimulation in Parkinson’s disease; a hierarchical Bayesian modeling of the ankle reaction time informs on the direct connection between the motor cortex and a muscle that controls the ankle dorsiflexion; the speed-accuracy tradeoff becomes an error signal in a simple control law to keep the children engaged during a robotic sensorimotor therapy for cerebral palsy.

At the forefront of this research, the interplay between behavior and neural activity is explored: From electrophysiological recordings of neurons to computational methods that emulate neural processes, principles and processes are extracted to form the very basis of what we experience as behavior and its pathology. Behind this research, there are human stories of how brain data science, behavioral psychology and clinical practice can amalgamate together and bridge the gap between disability and ability, between human limitation and human potential.