Changes in neural readout of response magnitude during auditory streaming do not correlate with behavioral choice in the auditory cortex

Abstract

Humans and non-human animals parse the auditory scene into distinct auditory objects or "streams" by grouping together stimuli with common features and segregating those with different features. This process is commonly called "auditory scene analysis". Although previous studies have identified neural mechanisms in the primary (core) auditory cortex that may underlie auditory-stream segregation, we do not have a good understanding of the contribution of cortical regions outside of the core auditory cortex to stream segregation nor do we understand the laminar specificity of these contributions. To examine these issues, we recorded translaminar multiunit activity (MUA) from the core and belt auditory cortex in macaque monkeys while they participated in an auditory streaming task designed to provide an objective measure of auditory-stream segregation. We found that MUA encoded both the stimulus variables and the monkey's behavioral choices related to our streaming task. Overall, core MUA was modulated more by the stimulus variables and the monkey's choices than belt MUA. However, whereas neural correlates of these variables were uniformly distributed in the core auditory cortex, stimulus- and choice-related information was predominantly represented in the superficial and deep layers of belt, respectively. These findings support a differential representation of stimulus- and choice-related variables related to auditory scene analysis along the primate auditory cortical pathway.

Laminar Neural Dynamics of Auditory Evoked Responses: Computational Modeling of Local Field Potentials in Auditory Cortex of Non-Human Primates

“…Moreover, the interpretation of activity (especially CSD) at later latencies is more uncertain due to the involvement of long-range cortical inputs (Happel et al, 2010). So far, evoked CSD and MUA have been extensively used to characterize tuning curves (Fishman et al, 2000a, 2000b; Fishman & Steinschneider, 2006, 2009; Steinschneider et al, 1998), differentiate responses to different stimuli (e.g., best-frequency [BF] vs. non-BF and standard vs. deviant) (Fishman & Steinschneider, 2012; Lakatos et al, 2020; O’Connell et al, 2011; Schaefer et al, 2015), and compare local neural population responses in different brain regions (e.g., core vs. belt regions of auditory cortex) (Banno et al, 2022). Statistical analyses in these studies mostly focused on neuronal activity occurring at specific latencies (usually within 50ms) in specific cortical layers (usually layer 4 or 2/3).…”