Diverse functions of the auditory cortico-collicular pathway

Abstract

Sensory processing is frequently conceptualized as a linear flow of information from peripheral receptors through hierarchically organized brain regions, ultimately reaching the cortex. In reality, this ascending stream is accompanied by massive descending connections that cascade from the cortex toward more peripheral subcortical structures. In the central auditory system, these feedback connections influence information processing at virtually every level of the pathway, including the thalamus, midbrain, and brainstem, and exert influence even at the level of the cochlea. The auditory cortico-collicular system, which connects the auditory cortex to the auditory midbrain, mediates manifold functions ranging from tuning shifts to defense behavior. In this review, we first summarize recent findings regarding the anatomical organization and physiological properties of the auditory cortico-collicular pathway. We then highlight several new studies that show that this projection system mediates high-level cognitive processes, acoustico-motor behaviors, and auditory plasticity, and discuss the circuit mechanisms through which they are mediated. Finally, we discuss remaining unanswered questions regarding cortico-collicular circuitry and function and potential avenues for future exploration.

Keywords: Auditory cortex; Cortico-collicular; Cortico-fugal; Inferior colliculus; Top-down.

Figure 1:. Summary of recent findings regarding the anatomy and physiology of the auditory CC pathway.

Layer 5 CC neurons are pyramidal cells that can generate calcium-dependent rhythmic bursts, while layer 6 CC neurons are regular spiking, do not burst, and display heavily-branched, horizontally-oriented dendritic arbors (orange and purple neurons). Layer 5 CC neurons receive columnar input from layer 5 and all cortical layers superficial to it (orange region). Layer 6 CC neurons receive input primarily from layer 6 itself (purple region). All layer 5 CC neurons receive monosynaptic excitatory thalamic input, while only 45% of layer 6 CC neurons receive direct thalamic inputs (blue arrows). Auditory inputs from the AC and the central nucleus of the IC (teal regions) interdigitate with neurochemical modules (yellow regions). In contrast, inputs from the somatosensory cortex and somatosensory brainstem regions directly target the neurochemical modules (red arrows to yellow regions). Ascending sound-evoked signals and CC feedback reach IC neurons with potential temporal overlap (green and orange/purple striped arrows). Descending signals can non-linearly amplify sound-driven responses in shell IC via NMDA-receptor dependent mechanisms (gray box).

Figure 2:. The CC pathway mediates attentional modulation of speech-in-noise perception.

Functional connectivity between primary AC (PAC) and brainstem (BS) was measured during A) passive and active speech-in-noise tasks on which the B) signal-to-noise ratio (SNR) was varied. C) Both bottom-up (left) and top-down (right) connectivity was enhanced during active compared to passive states. CC top-down connectivity decreased in noise, but only during passive states, suggesting that attention maintains top-down signaling in challenging listening conditions. Adapted from Price et al. 2021.

Figure 3:. CC input shapes predictive coding in IC.

A) All tones in the “cascade” sequence are presented with the same likelihood and no tone is repeated on adjacent tone presentations. Therefore, the neuronal response to a tone in this sequence (orange) can be compared to the response to the same tone when it is presented as a standard (gray) to compute repetition effects. Similarly, prediction effects can be computed by comparing the response to a tone in the cascade sequence to the response to the same tone when it is presented as an unpredictable deviant (blue). B) Prediction error is measured as a higher response to a deviant context than the cascade context, while the reverse relationship represents negative prediction error. A higher response to the cascade context than the standard context indicates repetition suppression, while a higher response to the standard context than the cascade context signifies repetition enhancement. C) The AC sends information about prediction error, negative prediction error, and repetition enhancement to the IC via the CC pathway. Repetition suppression was unaffected by CC suppression, suggesting that it may be a bottom-up process.

Figure 4:. CC input is implicated in gap detection.

Gap responses in an example IC neuron for control trials (black) and trials on which the AC was optogenetically suppressed (blue). During suppression, gap termination responses for brief gaps (4–32 ms) were suppressed, but longer gaps were unaffected, suggesting that responses to brief gaps are amplified in AC and routed subcortically. Adapted from Weible et al. 2020.

Figure 5:. The role of the CC system in plasticity following hearing loss.

A) Auditory nerve and CC growth functions were measured using auditory brainstem responses (ABRs) and two photon imaging of CC axon terminals, respectively. B) Imaging was performed daily for 7-14 days after noise exposure, and ABRs were measured every other day. C) Auditory nerve growth functions plummet after noise exposure (Day 1) and do not recover to baseline values on subsequent days, whereas D) CC growth functions supersede baseline values by Day 2 and remain persistently elevated. CCol = CC.