Examples of auditory responses to normal (forward) bird's own song (BOS). (A) Temporally reversed BOS (reverse BOS; B), and BOS in which syllable order was reversed [reverse order BOS (roBOS); C], in a single Area X neuron from a normally raised bird. (A–C) Each left panel shows sonograms (top; plots of frequency vs time, with amplitude indicated by the darkness of the trace) and oscillograms (bottom; plots of time-varying sound pressure) of the song stimulus. Each syllable is identified by a letter on the top; two motifs (‘a’–‘e’) and preceding introductory notes (‘i’) are shown. Note that the roBOS maintains the temporal order within individual syllables but reverses the syllable sequence in a song. Each right panel shows a PSTH of neural activity (30-ms bins; top) and an oscillogram of the song (bottom) presented in the left panel; 15 trials of each song were presented. Note that roBOS evoked responses weaker than BOS but stronger than reverse BOS.

Syllable order selectivity of Area X neurons in normal birds. (A) The mean response strength (RS) to BOS and roBOS in individual neurons are plotted. The diagonal line indicates where cells would lie if the mean values of RS to the two stimuli were equal. (B) Syllable order selectivity was quantified with a selectivity index (SI) calculated from the mean RS to BOS and roBOS in individual neurons (see Materials and methods). SI values of individual neurons are distributed widely from 0.5 (non-selective) to 1.0 (highly selective), indicating highly variable syllable order selectivity across different neurons.

Highly variable degrees of syllable order selectivity in Area X neurons. (A) Single-unit recordings of an Area X neuron with low syllable order selectivity from one bird, ‘normal 15’ (top), and of an Area X neuron with high syllable order selectivity from a different bird, ‘normal 16’ (bottom). The PSTHs show the neural responses to 20 presentations of forward BOS and roBOS. The normal 15 neuron had responses to roBOS that were as strong as that to BOS, whereas the normal 16 neuron had weak responses to roBOS compared with BOS. (B) The degree of syllable order selectivity in Area X neurons recorded from individual birds including normal 15 and normal 16. The circles each illustrate a neuron; each number on the abscissa refers to an individual bird; dotted horizontal line represents selectivity index (SI) = 0.5 (no order selectivity). Only data from birds in which more than two neurons were recorded are shown. The degree of syllable order selectivity is significantly more variable across different birds than across different neurons within individual birds (P < 0.0001, one-way anova across different birds).

Scatter plot compares syllable order selectivity with mean syllable duration of song in Area X neurons (top); oscillograms of two example songs are shown below, one with relatively long (middle) and one with short (bottom) mean syllable duration. The dots each illustrate a neuron; lines indicate the linear least squares fit though the data.

Possible neural coding mechanisms underlying syllable order selectivity. In each row, schematic drawings of simplified amplitude envelopes of BOS (left column) and its roBOS (middle column) are shown on the bottoms, and the expected neural responses (PSTHs) to those stimuli are shown on the top. The degree of syllable order selectivity obtained from the neural responses is indicated in the right column. BOS stimulus starts with three introductory notes, which are followed by simplified syllables labeled with lower case letters. (A) A model with a neural integration time window with fixed duration. Dotted rectangles represent integration time windows with fixed duration. Note that a song with syllables longer than the neural integration time window leads to low syllable order selectivity (top) and vice versa (bottom), while song with a mixture of syllable lengths is likely to have intermediate order selectivity (middle). (B) A song that contains similar syllables can also lead to low order selectivity, regardless of neural integration time.

Syllable order selectivity with different syllable definitions. (A) Oscillogram of a song of a bird ‘normal 17’. Introductory notes and the subsequent two motifs (dashed rectangles) are shown. (B) Two different segmentations of the first motif of the song. Sonogram (top) and oscillogram (middle) of the first motif are shown on an expanded time scale. The dashed lines indicate the boundaries of song segments defined as points where the song's envelope crosses a threshold, which was set as five times the RMS amplitude of the background noise. The silent intervals shorter than 2 ms were discarded, and the notes before and after the interval were combined into one syllable. These syllables defined in this way are labeled as ‘a’–‘i’ in the bottom. The syllables labeled ‘A’–‘E’ were defined by discarding longer intervals (< 20 ms). (C) Auditory responses of a single Area X neuron to BOS and roBOS with different syllable definitions. This neuron showed strong responses to the roBOS that was made using syllables ‘A’–‘E’ (roBOS2), as well as to forward BOS, while it responded much more weakly to the roBOS made by syllables ‘a’–‘i’ (roBOS1). The song oscillograms and the syllable labels are shown underneath the PSTHs; 20 trials of each song were presented. (D) The selectivity indices (SI) of all the Area X neurons recorded from three birds, calculated from the responses to BOS and the two different versions of roBOS (roBOS1 and roBOS2) created using the same approach as shown in (C). Two circles connected with a line illustrate each neuron; filled circles with a thick line represent the neuron whose responses are shown in (C). Seventeen out of 19 neurons showed a decrease in syllable order selectivity when mean syllable duration was artificially increased by changing syllable definition.

Acoustic similarity between forward BOS and roBOS measured using sliding windows is positively correlated with mean syllable duration. (A) Possible relationships between syllable order selectivity, acoustic similarity between BOS and roBOS, and mean syllable duration. The relationship indicated by arrow 2 will be examined in (B) and (D), and that by arrow 1 in Fig. 8. (B) Schematic showing the calculation of the average BOS–roBOS acoustic similarity in a bird ‘normal 14’. The acoustic similarities were obtained from comparisons of forward BOS and roBOS spectrograms using time windows of fixed duration sliding through both songs (red rectangles on the spectrograms, see Materials and methods). Example time windows (100 ms duration) on roBOS and their best matching windows on BOS are shown with the cross-correlation coefficients; solid and dashed arrows indicate strong and weaker correlations, respectively. To find the window size that best explains neural selectivity, the acoustic similarity measure was calculated multiple times, each time with a different, fixed acoustic comparison window size, ranging from 10 ms to 500 ms, and then compared with syllable order selectivity of Area X neurons (see Fig. 8). (C) The average BOS–roBOS acoustic similarities calculated with 10-, 100- and 500-ms windows (green, red and blue dots) are plotted against mean syllable duration of individual birds. The dots each illustrate a bird. The lines indicate the linear least squares fit though the data with each window size; the correlation coefficients (r) are shown for each analysis. (D) Correlation coefficients (r) between average BOS–roBOS acoustic similarity and mean syllable duration plotted as a function of the window size used to estimate the similarity. The filled circles indicate those correlations that are statistically significant (P < 0.05). Note that the acoustic similarity measure calculated with 50–200-ms windows was significantly correlated with mean syllable duration.

Average BOS–roBOS acoustic similarity can account for syllable order selectivity. (A) The average BOS–roBOS acoustic similarities calculated with 10-, 100- and 500-ms windows (green, red and blue dots) are plotted against the syllable order selectivity (SI) of individual Area X neurons. The dots each illustrate a neuron. Lines indicate the linear least squares fit though the data, and the correlation coefficient (r) is shown for each analysis. (B) Correlation coefficient (r) between syllable order selectivity (SI) and average BOS–roBOS acoustic similarity plotted as a function of the window size used to estimate the similarity. The filled circles indicate those correlations that are statistically significant (P < 0.05). Note that the acoustic similarity measure was significantly and negatively correlated with syllable order selectivity when the similarity measure was calculated with intermediate window durations.

Area X neural properties in isolate birds, revealed with the same analyses that were used in normal birds, are comparable to those in normal birds. (A) The mean response strength (RS) to BOS and roBOS in all Area X neurons recorded from isolate birds. Conventions are as in Fig. 2A. (B) Highly variable selectivity index (SI) values in isolates' Area X neurons. Conventions are as in Fig. 2B. (C) Negative correlation between syllable order selectivity and mean syllable duration of individual songs. Conventions are as in Fig. 4. (D) Decrease in syllable order selectivity after artificial increase of mean syllable duration by changing syllable definitions using the same approach as shown in Fig. 6. Conventions are as in Fig. 6D. (E) Correlation coefficient (r) between average BOS–roBOS acoustic similarity and mean syllable duration plotted as a function of the window size used to estimate the similarity. The filled circles indicate those correlations that are statistically significant (P < 0.05). Conventions are as in Fig. 7D. (F) Correlation coefficients (r) between syllable order selectivity and average BOS–roBOS acoustic similarity are plotted as a function of the window size used to estimate the similarity. The filled circles indicate those correlations that are statistically significant (P < 0.05). Conventions are as in Fig. 8B.