Sensory error drives fine motor adjustment

Abstract

Fine audiovocal control is a hallmark of human speech production and depends on precisely coordinated muscle activity guided by sensory feedback. Little is known about shared audiovocal mechanisms between humans and other mammals. We hypothesized that real-time audiovocal control in bat echolocation uses the same computational principles as human speech. To test the prediction of this hypothesis, we applied state feedback control (SFC) theory to the analysis of call frequency adjustments in the echolocating bat, Hipposideros armiger . This model organism exhibits well-developed audiovocal control to sense its surroundings via echolocation. Our experimental paradigm was analogous to one implemented in human subjects. We measured the bats’ vocal responses to spectrally altered echolocation calls. Individual bats exhibited highly distinct patterns of vocal compensation to these altered calls. Our findings mirror typical observations of speech control in humans listening to spectrally altered speech. Using mathematical modeling, we determined that the same computational principles of SFC apply to bat echolocation and human speech, confirming the prediction of our hypothesis.

Perceptual hearing sensitivity during vocal production

“…They not only dynamically adjust the structure of echolocation signals to ongoing navigation and foraging tasks 37 , 38 , 39 but also rapidly modify signal features to cope with interfering sounds. 40 , 41 , 42 , 43 , 44 By contrast, little is known regarding the ability to adaptively adjust signal parameters at a longer timescale, which is relevant to long-term learning. 45 , 46 We suggest that the slow vocal adaptation to auditory masking in H .…”

Doppler shift compensation performance in Hipposideros pratti across experimental paradigms

“…transported in a moving pendulum setup (Figure 3, bottom panels). In a detailed study with the same playback setup on another Hipposiderid bat, H. armiger, we have shown that online vocal frequency adjustments by H. armiger are driven by sensory prediction errors, but not by DSC or by jamming avoidance response (Wang et al, 2022). Thus, below we only made detailed statistical comparisons of the DSC performance of H. pratti in the free flight and moving pendulum paradigms.…”