Bored birds, busy brains: Habituation to song initiates significant molecular changes in auditory forebrain of zebra finch

When we think of habituation, we tend to think of a process in which there is a decrease in psychological and behavioural response(s) over time following an organism’s exposure to a stimulus. Conceptualising habituation in this manner seems to imply the loss of something once an initial learning event has taken place. Although this may accurately describe what occurs at the psychological and behavioural levels, a study by a group of scientists from the University of Illinois (Dong et al. 2010), which examines habituation at the neurobiological level, shows that contrary to this conceptualisation, both initial exposure and habituation to song playbacks initiates a vast array of genetic activity in the zebra finch brain.

The systematic regulation of FoxP2 expression in singing zebra finches has been the subject of previous posts, but there is also a growing literature, of which Dong et al’s study is a part, documenting increases in ZENK gene (which encodes a transcription factor protein that in turn regulates the expression of other target genes) expression in zebra finch auditory forebrain areas in response to playbacks of song or the song of a conspecific. Studies showed that ZENK expression seems to mirror the typical decline in response associated with habituation in that after a certain amount of repetition, presentation of the song that originally elicited upregulation of ZENK no longer did so, and that ZENK returned to baseline levels - although upregulation of ZENK would occur if a different song or an aspect of novelty was introduced (i.e. the original song was presented in a different visual or spatial context).

What Dong et al. have demonstrated by conducting a large scale analysis of gene expression at initial exposure, habituation, and post-habituation stages however, is that unexpectedly profound genetic changes occur as a result of habituation in the absence of any additional novel stimuli following the surge of activity observed during initial exposure to novel song. Thus, the resounding merits of the Dong et al. (2010) study lie in the broadness of their approach, providing a true sense of magnitude with respect to genomic involvement in vocal communication and illuminating important influences that have gone unnoticed by studies with a narrower focus. I summarise the experimental design and findings of the paper below.

Dong et al. 2010 - Experimental Design

The researchers measured both RNA and protein expression in the Auditory Lobule or AL (a region of the zebra finch brain that previous experiments have flagged for increased ZENK mRNA expression in response to song exposure) of zebra finches. Using an experimental design based on 3 main conditions over a time period of 3 days, subjects were distributed into the groups below. Half the birds in each group were euthanized for dissection of the AL so that RNA (including ZENK and related RNAs) could be extracted, where as the other half were euthanized an hour later so that protein expression could be analysed.

Three main groups used in initial experimental design:

1)       Silence – Subjects in this group did not hear any song what so ever over the three day period, providing a benchmark with which to compare the results of novel and habituated groups.

2)       Novel – Subjects in this group were exposed to novel song for the first time on the third day.

3)       Habituated – Subjects in this group were exposed to novel song for 3hrs on the second day (habituation) and heard a repeat of this song on the third day (a day after habituation had taken place).

Dong et al. 2010 - Results and findings

Changes in RNA regulation and protein expression in novel and habituated groups

Using microarray analysis, the team focused on RNA expression for each group and how this differed relative to the silence group. The sheer number of genetic components they found to be involved is astounding – 616 RNAs showed significant differences in the novel group relative to the silence group. Most significantly and rather unexpectedly however, was the whopping 2,923 additional RNAs that showed significant differences in birds in the habituated condition relative to silence that did not manifest in the novel group!

The use of 2-dimensional difference gel electrophoresis to analyse changes in protein expression also confirmed that mirroring RNA changes, proteins also displayed differential abundance (both up and down) relative to the silence group with additional protein changes in the habituated group. The results gained from both these approaches suggest different molecular profiles for each ‘state’, including a surprisingly distinct profile for birds in the habituated group relative to those in other groups.

Distinct molecular profile in habituated group is a ‘delayed’ consequence of training

To further investigate the nature of the distinct molecular profile observed in the habituated group, researchers added another condition to their experimental design, namely the ‘Trained only’ group. Birds in this group experienced the same exposure to song as those in the habituated group on the second day (3hr period), but did not that hear the song again on the third day. As molecular profiles were identical for birds in both habituated and trained only groups despite the lack of repeated exposure in the latter group, Dong et al. conclude that the distinct molecular profile of the habituated group is a ‘delayed’ molecular consequence of the previous days of training and was not a response to the re-presentation of familiar song.

Gene Grouping and Functional Analysis

The next step for the team was to try and get an idea of the corresponding function of these apparently significant and extensive molecular changes. Functional information was derived from genetic orthologs in the chicken, the only bird to have an entirely sequenced genome at the time this paper was published, which led the researchers to conduct a broader analysis of gene groups rather than detailed analyses of individual genes due to incomplete genetic data for the zebra finch. Functional analysis began with the classification of regulated RNAs into four groups according to the nature of their expression (up or down) relative to silence group (as shown below). Interestingly, RNAs were shown to have similar functional associations unique to their group.

Novel up (153 RNAs) and Novel down (463 RNAs) As expected, ZENK was identified as being among RNAs that showed the most significant increase relative to ‘Silence’ group, showing consistency with results of previous studies. Like ZENK, most of these RNAs encoded transcription factor proteins that regulate other genes. As reflected in the above figures, downregulated RNAs far outnumbered those upregulated RNAs present at the time of exposure constituting the initial surge of ‘positive’ activity. Identifiable RNAs appeared to be ion channel RNAs, however, the team speculated that a large proportion of RNAs that could not be identified for function represent noncoding RNAs.

Habituated up (1,467 RNAs) and Habituated down (RNAs 1,456) Despite the vast number of differentially expressed gene products apparent in the habituated group, the team were able to discern changes in a high majority of RNAs and proteins involved in mitochondrial regulation (mitochondria being known as the ‘power houses’ of cells), suggesting the onset of a widespread metabolic or energetic shift in the AL as a consequence of the habituation process.

Looking forward

So, it goes to show that what looks to be happening on the ‘outside’ doesn’t necessarily reflect what is happening on the ‘inside’.  Even if birds seem a bit fed up after being subjected to the same pesky song over and over again (who wouldn’t!), their genomes are hubs of activity. In a way this isn’t surprising in that habituation is a form of learning (albeit a basic one) in which knowledge and information must be integrated and neural circuits modified. For further detail and clarification I strongly encourage a reading of Dong et al.’s paper, which is brilliantly written and fairly accessible to non-experts in biology such as myself. It is also essential work that has provided a foundation for coming research on the recently assembled zebra finch genome (hoping to post about this soon), which will no doubt yield some fascinating insights.


Dong, S et al. (2009) Discrete molecular states in the brain accompany changing responses to a vocal signal. PNAS 106:27, 11364-11369