Starts 8 Nov 2016 11:00
Ends 8 Nov 2016 12:00
Central European Time
Distinct timescales of population coding across cortex
ICTP
Central Area, 2nd floor, old SISSA building
Abstract:
The cortex represents information across widely varying timescales. For instance, sensory cortex encodes stimuli that fluctuate over milliseconds, whereas in association cortex behavioural choices can require the maintenance of information over seconds. It is poorly understood how the cortex achieves such diverse timescales of information coding. While recent work has identified different timescales in features intrinsic to individual neurons, the timescales of information coding in populations of neurons have not been studied, and population codes have not been compared in depth across cortical regions. In this talk, I will compare coding for sensory stimuli and behavioural choices in auditory cortex (AC) and posterior parietal cortex (PPC) in the mouse brain, as the animal performs a sound localisation task. I will show that population codes are essential to achieve long and diverse coding timescales, and that the statistical structure of the codes differs between sensory and association cortices.
Among PPC neurons, correlations that are not explained by task events ("functional coupling") extend over long time lags, and contribute to a long timescale population code characterised by consistent representations of choice lasting over two seconds. In contrast, coupling among AC neurons is weak, shorter-lived, and results in moment-to-moment fluctuations in stimulus and choice information. This suggests that population coupling is a variable property that affects the timescale of information coding: relatively uncoupled activity in sensory cortex is key for signals that change rapidly to code temporally variable stimuli, whereas highly coupled activity in association cortex appears critical to form a consistent signal from which temporally integrated information can be read out instantaneously to drive behaviour.
The cortex represents information across widely varying timescales. For instance, sensory cortex encodes stimuli that fluctuate over milliseconds, whereas in association cortex behavioural choices can require the maintenance of information over seconds. It is poorly understood how the cortex achieves such diverse timescales of information coding. While recent work has identified different timescales in features intrinsic to individual neurons, the timescales of information coding in populations of neurons have not been studied, and population codes have not been compared in depth across cortical regions. In this talk, I will compare coding for sensory stimuli and behavioural choices in auditory cortex (AC) and posterior parietal cortex (PPC) in the mouse brain, as the animal performs a sound localisation task. I will show that population codes are essential to achieve long and diverse coding timescales, and that the statistical structure of the codes differs between sensory and association cortices.
Among PPC neurons, correlations that are not explained by task events ("functional coupling") extend over long time lags, and contribute to a long timescale population code characterised by consistent representations of choice lasting over two seconds. In contrast, coupling among AC neurons is weak, shorter-lived, and results in moment-to-moment fluctuations in stimulus and choice information. This suggests that population coupling is a variable property that affects the timescale of information coding: relatively uncoupled activity in sensory cortex is key for signals that change rapidly to code temporally variable stimuli, whereas highly coupled activity in association cortex appears critical to form a consistent signal from which temporally integrated information can be read out instantaneously to drive behaviour.