Referee 3's review:

In this study, Jarvis and colleagues propose a model that is intended to explain the similarities of vocal control circuits of vocal learning bird species, the songbirds, parrots and hummingbirds. For each of these taxa, the authors propose that the forebrain vocal control system is composed of seven homologous regions. The authors hypothesize that there is a preexisting anatomical constraint (due to common evolutionary history) that determined evolution of forebrain areas involved in song learning in the three avian taxa. As this constraint, the authors suggest a basic avian pattern of brain areas that control activity of limb and body movements. They then suggest that regions involved in vocal learning and the control of learned vocal pattern have evolved from this motor (movement-associated) pathway(s). The data that indicate movement-associated areas are in general o.k. (but see below). The main finding is the IEG expression in 11 forebrain regions, 7 of these areas are thought to be motor (movement associated) and 4 such areas are thought to be somatosensory. The 7 motor only regions are then thought to be adjacent to forebrain vocal control nuclei. This relation is, however, highly speculative and not supported by the data. In a way, many things are adjacent in a small brain.

As a technical approach, the authors map the expression of immediate early genes (IEG) following vocalizations and other movements in relation to visual and auditory input. The IEG-labeling and behavioral methods used are in general sound except that the authors have no real control for the presence/absence of somatosensory input. Further, the movement-related expression of IEG is in some cases such as the hummingbird forebrain or the songbird rostral forebrain rather widespread. The electrophysiology part is incomprehensive due to the lack of procedural details.

Concerning the data analysis, there is a major technical problem in that the authors have no measurement for “adjacent”, i.e. there are no quantifications of the anatomical distances between song areas and “nearby” movement-associated areas.
Due to the pattern of widespread movement-associated expression of IEG and the lack of adequate distance data between the various brain regions we have the following ambiguous situation: It seems that many region in the rostral forebrain (except MAN and Area X) are IEG-labeled (Fig. 3) and that some areas in the caudal forebrain are labeled with little label in the arcopallium, although the authors indicate labeling (LAI, adjacent to RA p.5) there. Further, PLN/PLMV which might be dorsolateral of song nuclei Nif and Av are not clearly visible in Fig. 3Af,h or appear different from those regions named PLN/PLMV in Fig. 2A. In relation, the authors state (last sentence, 1st paragraph, p. 5) that it was difficult to determine if the “movement (wing beats, flights)” regions were adjacent to vocal nuclei. In particular, areas thought to be “near” HVC and RA (DLN and LAI) are difficult to detect in IEG labeling or appear rather distant from HVC and RA. Further, IEG expression in PLN and PLMV appear to be both, auditory and motor, questioning further the conclusion that song areas are (evolutionary) derived from general motor areas. Further, movement-associated IEG areas that are in the view of the authors not adjacent to vocal forebrain regions are classified as somatosensory. However, this does not exclude that they are in part motor, like PLN/PLMV are auditory and movement-associated.

Concerning the basic concept, it is unclear to the reviewer how the authors deduce the number of 7 forebrain vocal areas. Obviously this depends on how one “counts”, i.e. what the definition of a nucleus is. E.g., the authors unify lMAN and mMAN as one nucleus despite very different projection pattern of these areas in songbirds. Further, it remains to be seen how similar forebrain region that might be involved in song control of parrots, hummingbirds, and songbirds are. In relation, the taxa comparisons, in particular that of songbirds and hummingbirds does not support the ideas of the authors due to very widespread expression of ZENK in moving hummingbirds. Further, as in his previous hummingbird paper (Jarvis et al., 2000), the PI cannot exclude that the areas that he determined as vocal areas in the hummingbird brain are involved in other functions since the animals were engaged in different behaviors (singing and “other” behaviors). Likewise, the expression pattern in non-learning pigeon in relation to movement is rather scattered, particular in the caudal forebrain where one would expect HVC and RA in songbirds.

Conceptually, the authors point to the fact that human language and song of some avian groups are similar in that they depend on auditory-motor learning. However, the first sentence of the abstract suggests that bird song and spoken human language are more similar than they are. Similarly, the last sentence of the introduction is non-sense, deducing a motor origin of spoken language from a motor origin of bird song learning. This statement claims that the hypothesized motor constraints of vocal learning were already present in a common ancestor of birds and mammals.

Last two conceptual point: The type of movements controlled by the indicated brain regions concern movements that do not involve much if any learning while song learning is a learned behavior. Further, these movements involve somatosensory feedback while song control requires auditory feedback; in relation there is as yet no evidence for spindles in syrinx muscles.

Minor points:

The added value of the FoxP1 expression experiment is not clear to the reviewer.

The first experiment indicating a role of the forebrain pathway in song learning is from Bottjer et al., 1984, not from references 8 and 9.

For the interpretation of the lateralized IEG expression in relation to reduced sensory input of hovering animals, it would be necessary to see the degree the lateralization in “singing/other behavior” hummingbirds. Unfortunately the authors do not show this information (Fig. 10)

Fig.1 & Fig. 12: The scheme of the hummingbird brain includes connectivity data that are not included in the previous publications to which the legend refers.

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.