Understanding Variability in Brain and Behavior
Variability is a pervasive feature of neural activity and behavior. Several theories postulate that variability is fundamental for adaptive brain function. I propose that understanding the sources and functional role of neural variability requires partitioning the variability of observed behavior. One partition is, in fact, only “apparent variability”: it is caused by dynamic variations in systematic decision biases (not decision noise), which often remain hidden. Only the second partition is “genuine variability”, which results from the inherent noisiness of cortical neurons. To support this claim, I will present our recent work on perceptual decision-making under uncertainty. We find that a large portion of the variability in perceptual choice behavior can be explained by dynamic trial-to-trial modulations of decision biases: apparent variability. Those bias modulations are induced by the history of previous choices as well as by phasic arousal transients during decision formation. They are implemented by systematic modulations of the underlying cortical activity states. In another line of work, we show that the level of genuine variability in perceptual choice is controlled by a major neuromodulatory system – the norepinephrine system – likely through altering the balance between excitation and inhibition in cortical networks. This line of work points to an adaptive function of noise in perceptual inference. I conclude that pinpointing the factors governing apparent behavioral variability and identifying control mechanisms for genuine behavioral variability will enable a deeper understanding of brain function. Key references: • Urai AE, Braun A & Donner TH. 2017. Pupil-linked arousal is driven by decision uncertainty and alters serial choice bias. Nature Communications. 8:14637. • de Gee JW, Colizoli O, Kloosterman NA, Knapen T, Nieuwenhuis S & Donner TH. 2017. Dynamic modulation of decision biases by brainstem arousal systems. eLife 6. pii: e23232. • Braun A, Urai AE & Donner TH. 2018. Adaptive history biases result from confidence-weighted accumulation of past choices. Journal of Neuroscience. pii: 2189-17. • Pfeffer T, Avramiea AE, Nolte G, Engel AK, Linkenkaer-Hansen K & Donner TH. 2018. Catecholamines alter the intrinsic variability of cortical population activity and perception. PLoS Biology. 16: e2003453. • Talluri BC, Urai AE, Tsetsos K, Usher M & Donner TH. 2018. Confirmation bias through selective over-weighting of consistent information. Current Biology. 28: 3128-3135.