Investigating the neural mechanisms of feature-based and spatial attention in a network model of two coupled brain areas
Zusammenfassung der Projektergebnisse
A key question of systems neuroscience is how the brain processes incoming sensory stimuli, giving rise to perception and action. The neural basis of the computations that underlie cognitive function has been studied extensively with decision making and working memory tasks in humans and monkeys. In this project, I have developed computational models of neuronal circuits to pin down the physiological mechanisms underlying an animal’s behavior during these elementary cognitive tasks. I have complemented the study of the dynamics of these circuits with quantitative analysis of psychophysical and electrophysiological data (single and multi-unit recordings, local field potentials). A tight interaction between theory and experiment, based on active collaborations with experimental groups has been the basis of this work (collaborators included C. Constantinidis, Wake Forest Univ.; T. Pasternak, Rochester Univ.). In particular, I investigated the relationship between trial-to-trial variability of neural activity (often referred to as “neuronal noise”) and behavioral responses, leading to the following main results of this project: (1) I have developed a hierarchical model of the interplay between bottom-up signals and top-down (attention) signals which prompts for a reformulation of the current standard framework of perceptual decision making. In this model, sensory neuronal variability is caused by bottom-up processes and top-down interactions across the hierarchy, thereby questioning its commonly assumed causal impact on behavior. I could show that the activity in MT neurons of monkeys performing a classical random dot discrimination task is consistent with several predictions derived from this model. (2) I have used a neural attractor model together with data from monkey experiments to reveal for the first time the neural basis of working memory precision in prefrontal cortex. Persistent activity in the delay periods of working memory tasks has long been hypothesized to be the neural substrate of working memory. This persistent activity has been speculated to be generated by continuous attractors (bump attractors) in the neuronal microcircuits of the prefrontal cortex, but this hypothesis lacked direct experimental validation. I demonstrated that trial-to-trial variations of persistent neural activity in the PFC of monkeys performing a spatial working memory task predict the pattern and magnitude of saccadic response inaccuracies exactly as expected from the bump attractor model. These results validate the idea that patterns of elevated PFC activity maintain spatial information through memory delays, and they reveal a hitherto unknown relationship between neural variability and behavioral variability in the delay period of the working memory task, long before behavior occurs. This work was featured on the cover of the March 2014 issue of Nature Neuroscience, it received two recommendations in F1000, and broad coverage in the media (Spanish newspapers La Razón, La Vanguardia; Spanish television: TV3).
Projektbezogene Publikationen (Auswahl)
- (2014) Bump attractor dynamics in prefrontal cortex explains behavioral precision in spatial working memory. Nat Neurosci 17:431–439
Wimmer K, Nykamp DQ, Constantinidis C, Compte A
(Siehe online unter https://doi.org/10.1038/nn.3645) - (2015) Sensory integration dynamics in a hierarchical network explains choice probabilities in cortical area MT. Nat Commun 6:6177
Wimmer K, Compte A, Roxin A, Peixoto D, Renart A, de la Rocha J
(Siehe online unter https://doi.org/10.1038/ncomms7177) - Transitions Between Multiband Oscillatory Patterns Characterize Memory-Guided Perceptual Decisions in Prefrontal Circuits. Journal of Neuroscience 13 January 2016, 36 (2) 489-505
Wimmer K, Ramon M, Pasternak T, Compte A
(Siehe online unter https://doi.org/10.1523/JNEUROSCI.3678-15.2016)
