PMC

Age-dependent sequence plasticity at branch point sequences. A: transition entropy significantly decreases over time in adult Bengalese finches. Plotted are transition entropies, a measure of sequence variability, of individual branch point sequences when birds were young adults (x-axis) and older adults (y-axis). Most of the points lie below the line of unity (dashed line), indicating that transition entropy was lower when birds were older adults than when they were younger adults (P < 0.001). B: many transitions were “pruned” (i.e., observed in the younger adult recordings but not in the older adult recordings), and most of these transitions were infrequently produced when birds were younger adults. Plotted is a distribution of transitions probabilities (as young adults) of transitions that were pruned with age. C: pruned transitions were produced less frequently than transitions that were retained over time. Plotted is a cumulative distribution of transition probabilities (Cum. Prob.; as young adults) of transitions that were pruned (dashed line) and retained (solid line) over time. D: the decrease in transition entropy was significantly greater for branch point sequences with at least 1 pruned transition than for branch point sequences without a pruned transition (*P < 0.05). The change in entropy was significantly different from 0 only for branch points with at least 1 pruned transition. E: the decrease in transition entropy was significantly greater for branch point sequences in which dominant transitions increased in prevalence than for branch point sequences in which dominant transitions decreased in prevalence (*P < 0.05). The change in entropy was significantly different from 0 only for branch points in which dominant transitions increased. F: the pruning of infrequently produced transitions and increases in the transition probabilities of dominant transitions independently contributed to the decrease in transition entropy with age. In a 2-way ANOVA with the presence of a pruned transition (yes or no) and the direction of change of the dominant transition (up or down) as independent factors, both transition pruning and increases in dominant transitions independently and additively contributed to a decrease in transition entropy over time.

Muscimol increased the variability of syllable sequencing at branch points, and the magnitude of effect varied depending on the inherent variability of syllable sequencing of the branch point. A: the change in transition entropy caused by muscimol infusions for each branch point (n = 22 branch points across 8 birds) in relation to the transition entropy of the branch point under control conditions. There is a significant negative relationship between the degree to which muscimol infusions increased transition entropy and transition entropy under control conditions (mixed-effects model). Each symbol represents a different bird. B and C: examples of 2 branch points from a single bird, 1 with low transition entropy under control conditions (B) and another with high transition entropy under control conditions (C). For each example, on left are spectrograms of the transitions following the branch point sequence (“defgh” in B and “abc” in C), and on right are the transition probabilities (%) for each of the branch point transitions following phosphate-buffered saline (PBS) or muscimol infusions. Also listed are the transition entropies under PBS and muscimol conditions. The degree of change in transition entropy is greater for the branch point in B than the branch point in C.

Reaction coordinates of a general reaction (schematic). (a) and (b) two dimensional projection of the conformational hypersurface, iso-energy lines are schematically shown. Migration pathways of molecules through configurational space shown as arrows. (a) Transition state # has low entropy; reaction proceeds through a narrow path, ΔS^# is negative. (b) Similar to (a), transition state # has a larger entropy, the molecules are more flexible at the transition state, ΔS^# is positive. (c) One dimensional projection of the configurational energy hypersurface with energy E_conf along the reaction coordinate. S₁^F: conformational entropy of state 1 in the flexible state, S₂^F: conformational entropy of state 2 in the flexible state, S^cs: conformational entropy of conformational substates. Here, S₁^F > S₂^F > S^cs and reaction entropy ΔS is negative. S^#: entropy of transition state. ΔH^#_conf: change of conformational enthalpy from state 1 to the transitions state.

Social context did not consistently modulate the degree to which feedback perturbations caused changes to syllable sequencing. A: example of changes to transition probabilities and entropy for the branch point sequence bcd (d is the target syllable). Spectrograms of the 3 possible transitions from the syllable d—defg (top), efg (middle), and i (bottom)—are plotted on the left, and the transition probabilities under control and altered feedback (AF) conditions for UD and FD song are listed on the right. Below the transition probabilities are the transition entropies for control and AF trials during UD and FD song (red text). To the right of the transition probabilities for both UD and FD song are listed the absolute differences in transition probabilities between control and AF trials (green text). For all experiments, we summarized the effect of feedback perturbation at branch points as 1) the percent change in entropy between control and altered feedback trials and 2) the mean (absolute) change in transition probabilities between control and altered feedback trials. For this example, feedback perturbation caused a smaller change in transition entropy (4 vs. 54%, respectively) and in transition probabilities (0.05 vs. 0.23, respectively) during FD song than during UD song. B: there was considerable variation across experiments in how feedback perturbation affected transition entropy across UD and FD songs. Consequently, there was not a significant difference between UD and FD song in the magnitude of effect of feedback perturbation on transition entropy. C: there also was considerable variation across experiments in how feedback perturbation affected transition probabilities across UD and FD songs, and consequently, there was not a significant difference between UD and FD song in the degree to which feedback perturbation affected transition probabilities. These data suggest that, in contrast to song interruptions and tempo, we did not detect a significant social modulation of the effect of auditory feedback perturbations on syllable sequencing.