Spatial and Temporal Organization of Composite Receptive Fields in the Songbird Auditory Forebrain

Abstract

The mechanisms underlying how single auditory neurons and neuron populations encode natural and acoustically complex vocal signals, such as human speech or bird songs, are not well understood. Classical models focus on individual neurons, whose spike rates vary systematically as a function of change in a small number of simple acoustic dimensions. However, neurons in the caudal medial nidopallium (NCM), an auditory forebrain region in songbirds that is analogous to the secondary auditory cortex in mammals, have composite receptive fields (CRFs) that comprise multiple acoustic features tied to both increases and decreases in firing rates. Here, we investigated the anatomical organization and temporal activation patterns of auditory CRFs in European starlings exposed to natural vocal communication signals (songs). We recorded extracellular electrophysiological responses to various bird songs at auditory NCM sites, including both single and multiple neurons, and we then applied a quadratic model to extract large sets of CRF features that were tied to excitatory and suppressive responses at each measurement site. We found that the superset of CRF features yielded spatially and temporally distributed, generalizable representations of a conspecific song. Individual sites responded to acoustically diverse features, as there was no discernable organization of features across anatomically ordered sites. The CRF features at each site yielded broad, temporally distributed responses that spanned the entire duration of many starling songs, which can last for 50 s or more. Based on these results, we estimated that a nearly complete representation of any conspecific song, regardless of length, can be obtained by evaluating populations as small as 100 neurons. We conclude that natural acoustic communication signals drive a distributed yet highly redundant representation across the songbird auditory forebrain, in which adjacent neurons contribute to the encoding of multiple diverse and time-varying spectro-temporal features.

Epi-Intra neural probes with glassy carbon microelectrodes help elucidate neural coding and stimulus encoding in 3D volume of tissue

“…In a recent study, we have demonstrated that CRFs of neurons are independent of the neurons' distance, location and, topology [28]. The stimuluslocked properties of recorded cells by the Epi-Intra probe proved to be useful in making them viable candidates for CRFs extraction.…”

“…As starters, to extract high quality composite receptive fields from neuron cell responses, the cell responses should be stimulus-locked with low noise [27,28]. Since the recorded cells by Epi-Intra probe exhibit such properties, as shown earlier, we were then able to extract CRFs from them quite easily.…”

“…Faced with this challenge of decreasing crosscorrelation of LFPs with distance, we, therefore, then proposed the investigation of an alternate approach involving composite receptive fields (CRFs) of neuron cells [27]. In a recent study, we already have shown that CRFs are independent of the neurons' location and their topology [28]. CRFs, therefore, appear, at least theoretically, to offer a potentially useful tool for investigating the usefullness of the Epi-Intra probe as a tool for neural coding over a larger brain area.…”

“…Ultimately, we were able to reconstruct bird song stimuli portions with this method by finding location of CRFs corresponding to specific portions of vocal elements of stimuli] (figure 9(b)). These methods are explained in detail elsewhere [27][28][29].…”

Predictive coding for natural vocal signals in the songbird auditory forebrain

“…To relate stimulus representations to neural activity, we computed composite receptive fields (CRFs) for auditory cortical neurons using the Maximum Noise Entropy (MNE) method (Fitzgerald et al, 2011). Previous work has shown that the MNE model well-describes the sensitivity to higher-order features exhibited by NCM neurons (Kozlov & Gentner, 2016;Vahidi, 2021).…”