Neuronal correlates of sensory working memory in the auditory cortex of humans an monkeys
Final Report Abstract
This project was an extension of a previous DFG-project, in which we identified cortical correlates of auditory working memory (WM) by means of experiments with humans and monkeys requiring them to perform a variety of tasks on sequences of two sounds separated by a silent delay on the order of seconds. These included tasks where a representation of the first sound had to be kept in WM during the delay so that it could be compared with the second sound. During task performance by humans, we acquired the electromagnetic activity of a large population of neurons using magnetoencephalography (MEG), a modern imaging method with high temporal resolution. During task performance by monkeys, we recorded the electrical activity of individual nerve cells and groups of cells in the auditory cortex (AC) by means of microelectrode arrays. In both species, we found that memory-related information was stored in AC for a short time. The storage was reflected in altered activity of nerve cells persisting throughout the period during which the first tone had to be memorized. With several control tasks and conditions, we could successfully exclude the possibility that the observed persistent activity merely reflects other auditory and/or cognitive processes. In the extension period, we combined our convergent approach with computational modelling of signal processing in AC using an adaptation-based model of the core-belt-parabelt organization of the primate AC. In additional experiments, , we observed that micro-stimulation of AC affected the monkeys' performance in WM tasks. These findings therefore provide evidence that the persistent activity in AC is causally linked to WM. Our modelling work provided deeper insight into potential mechanisms of signal processing in the AC in general and into the origin of sensory and WM-related persistent activity in AC in particular. Our model is able to replicate well the increase of the peak amplitude of the response as a function of stimulus-onset interval, usually characterized by a saturating exponential function with time constant 𝜏SOI . Our simulations predict that this adaptation time constant 𝜏SOI from intra- and extracranial measurements depends on the intrinsic recovery time constant of AC as well as on tone frequency and location of the underlying cortical generator. This implies that the interplay between synaptic plasticity, tonotopic representation, and the analysis of time information in cortex might form a richer tapestry than previously realized. Preliminary simulations of the high- and low-WM load conditions of our WM experiments suggest that excitatory input from PFC to parabelt may be required to generate a persistent field throughout the delay, as observed experimentally. Finally, we also sought analytical solutions to the non-linear model by linear approximations of the spiking rate and the adaptation term which result in a description of the AC dynamics in terms of the basic elements of the decoupled state variables, the socalled normal modes. So far, using a linearized spiking rate, we are on the way to unveil the link between the weight matrices of the model and the waveforms of stimulus-evoked MEG responses. Analytical solutions also allow one to find subject-specific model parameters by fitting simulated responses to event-related field data in a computationally fast and efficient way.
Publications
-
(2016) Neuronal activity in primate prefrontal cortex related to goaldirected behavior during auditory working memory tasks. Brain Res 1640 (Pt B): 314 – 327
Huang Y, Brosch M
-
(2016) Persistent neural activity in auditory cortex is related to auditory working memory in humans and nonhuman primates. eLife 2016;5:e15441
Huang Y, Matysiak A, Heil P, König R, Brosch M
-
(2017) Category-specific neuronal activity in the primate auditory forebrain. PLoS One 12(10):e0186556
Selezneva E, Oshurkova E, Scheich H, Brosch M
-
(2018) Neuronal correlates of auditory streaming in monkey auditory cortex for tone sequences without spectral differences. Frontiers in Neuroscience
Knyazeva S, Selezneva E, Gorkin A, Aggelopoulos NC, Brosch M
