A lobule-specific neuronal representation of song temporal structure in the songbird cerebellum

Reviewer #1 (Public review):

In this study, Ursu, Centeno, and Leblois record from the cerebellum of zebra finches and analyze neurons for auditory and song-related activity. The paper covers a lot of ground, ranging from lesions of the deep nuclei to song and white noise playback inside and outside of singing, and some level of survey of response types across cerebellar lobules, to provide foundational information on cerebellar relationships with song. There are a number of interesting observations in the study, to me most notably, the lack of responsivity of song-related activity in lobule IV to distorted auditory feedback. This observation is interesting in light of the perennial idea that the cerebellum may participate in rapid error corrections in other somatic control domains. If such a role were relevant for song, it stands to reason that some alteration of activity could be found there. Of course, on the other hand, zebra finches do not show rapid corrections during DAF, so perhaps the null result does not resolve much. Nevertheless, these data are important steps forward in establishing the involvement or lack of involvement in a broader set of brain structures beyond the song control system typically studied. While the study presents some interesting and important inroads, in my opinion, there was a general lack of 'polish' to the study that led to ambiguity in the report and confusing displays. This detracted from rigorous reporting of the findings.

Reviewer #2 (Public review):

In this paper, the authors investigate the role of the cerebellum in song production in the zebra finch. First, they replicate prior studies to show that lesions of the lateral deep cerebellar nuclei (latDCN, primarily lobules IV-VII and IX) result in shorter duration syllables and song motifs than sham controls. The authors then record neural activity from the cerebellum during both passive auditory exposure in anesthetized birds and in freely singing animals. The authors claim that across multiple lobules, the cerebellum receives "non-selective" auditory inputs locked to syllable boundaries (based on acute recordings) and that cerebellar neurons display song-locked responses that are unaffected by auditory feedback perturbations (in chronic recordings). Moreover, the authors emphasized the distinct properties of lobule IV, which they argue is tightly locked to the onset and offset of syllables, and conclude that the cerebellum might contribute to the duration of song elements.

This paper presents novel and useful descriptions of song-related neural activity in the cerebellum. However, there are multiple serious issues. First, there are major issues with the design and presentation of the analysis of the electrophysiological data; based on these, it is unclear whether the authors are justified in some of their conclusions about neural tuning or are entitled to any of their claims about the specific tuning or function of neurons in particular lobules. Second, because the authors' conceptual framework seems to ignore possible non-auditory inputs to the cerebellum, their results on (minimal) effects of auditory manipulation during singing are over-interpreted with respect to providing evidence of a forward model. Third, the paper's central assertion - that the songbird cerebellum may contribute to the duration of vocal events during song - was firmly established by a prior lesion study (Radic et al., 2024). Although the authors do cite this prior study with respect to longer-term postlesion changes after cerebellar lesions, this paper also showed a large change in syllable duration immediately after cerebellar lesion (Figure 5 in Radic et al). The electrophysiological results in the present paper could provide valuable insights into the neural mechanisms underlying this already-described role of the songbird cerebellum; however, given the other concerns above, it is not clear that the authors have done so.

Author response:

Public Reviews:

Reviewer #1 (Public review):

In this study, Ursu, Centeno, and Leblois record from the cerebellum of zebra finches and analyze neurons for auditory and song-related activity. The paper covers a lot of ground, ranging from lesions of the deep nuclei to song and white noise playback inside and outside of singing, and some level of survey of response types across cerebellar lobules, to provide foundational information on cerebellar relationships with song. There are a number of interesting observations in the study, to me most notably, the lack of responsivity of song-related activity in lobule IV to distorted auditory feedback. This observation is interesting in light of the perennial idea that the cerebellum may participate in rapid error corrections in other somatic control domains. If such a role were relevant for song, it stands to reason that some alteration of activity could be found there. Of course, on the other hand, zebra finches do not show rapid corrections during DAF, so perhaps the null result does not resolve much. Nevertheless, these data are important steps forward in establishing the involvement or lack of involvement in a broader set of brain structures beyond the song control system typically studied. While the study presents some interesting and important inroads, in my opinion, there was a general lack of 'polish' to the study that led to ambiguity in the report and confusing displays. This detracted from rigorous reporting of the findings.

We thank reviewer #1 for his comments. We will clarify the possible misleading or ambiguous claims and interpretations in the present manuscript and polish the presentation of the results. We will also modify the discussion to better place or results within the current knowledge on cerebellum and songbirds, and in particular address the link between our findings and the low sensitivity to auditory feedback in zebra finches.

Reviewer #2 (Public review):

In this paper, the authors investigate the role of the cerebellum in song production in the zebra finch. First, they replicate prior studies to show that lesions of the lateral deep cerebellar nuclei (latDCN, primarily lobules IV-VII and IX) result in shorter duration syllables and song motifs than sham controls. The authors then record neural activity from the cerebellum during both passive auditory exposure in anesthetized birds and in freely singing animals. The authors claim that across multiple lobules, the cerebellum receives "non-selective" auditory inputs locked to syllable boundaries (based on acute recordings) and that cerebellar neurons display song-locked responses that are unaffected by auditory feedback perturbations (in chronic recordings). Moreover, the authors emphasized the distinct properties of lobule IV, which they argue is tightly locked to the onset and offset of syllables, and conclude that the cerebellum might contribute to the duration of song elements.

This paper presents novel and useful descriptions of song-related neural activity in the cerebellum. However, there are multiple serious issues. First, there are major issues with the design and presentation of the analysis of the electrophysiological data; based on these, it is unclear whether the authors are justified in some of their conclusions about neural tuning or are entitled to any of their claims about the specific tuning or function of neurons in particular lobules. Second, because the authors' conceptual framework seems to ignore possible non-auditory inputs to the cerebellum, their results on (minimal) effects of auditory manipulation during singing are over-interpreted with respect to providing evidence of a forward model. Third, the paper's central assertion - that the songbird cerebellum may contribute to the duration of vocal events during song - was firmly established by a prior lesion study (Radic et al., 2024). Although the authors do cite this prior study with respect to longer-term postlesion changes after cerebellar lesions, this paper also showed a large change in syllable duration immediately after cerebellar lesion (Figure 5 in Radic et al). The electrophysiological results in the present paper could provide valuable insights into the neural mechanisms underlying this already-described role of the songbird cerebellum; however, given the other concerns above, it is not clear that the authors have done so.

We thank reviewer #2 for these comments. We will improve the presentation of the results, in particular our cell-type classification of the electrophysiology recordings based on latest literature and the statistics of the tuning differences between lobules. We will also modify the discussion regarding singing related internal models and consider non-auditory feedback. Finally, we will clarify the position of our work within the existing songbird literature and clarify what are the specific contributions of this work. We fully agree that prior studies have already shown the behavioural effects of lesions, as already clearly mentioned in introduction and discussion, and rather aimed at reproducing partially these results before diving into neural mechanisms. We will clarify this point in our revision.