Singing Mice Paper Flashcards
Adaptive behaviour
Often requires adjusting action in response to a rapidly changing environment.
Acoustic communication
Often requires rapid modification of motor output in response to sensory cues.
Problem being addressed
Little is known about mechanisms of sensorimotor transformations (their area of interest is in natural contexts, like social behavior)
Few models of acoustic interactions in natural world exist.
What is so interesting about social interactions
Animals must dynamically modulate complex actions in response to the changing behaviour of a conspecific.
Why focus on acoustic exchanges?
They are a promising foci for the study of sensorimotor transformations that underpin social behaviour. Very common across different taxa (insects, birds, amphibians, mammals). Require dynamic interactions since animals must avoid overlapping with one another.
Also, serves a variety of essential social functions: mate selection, male-male competition.
Why did they choose to study Alston’s singing mouse? (Scotinomys teguina)
They have robust and rapid countersigning (500ms) that resembles the sub second latencies of conditioned sensorymotor transformations in lab settings + the timing of vocal turn taking seen in human conversation.
Vs. Lab mice (which don’t demonstrate turn-taking behaviour even though they produce frequency-modulated vocalisations).
Vs. Marmoset pairs (which show antiphonal calling but it’s too slow).
Goal/how they solved problem
Examined vocal exchanges in Alston’s singing mouse (Scotinomys teguina).
Manipulated neural dynamics to determine motor cortical locus that works hierarchically within song production pathway to enable precise vocal interactions between pairs of mice
Essentially, what area of the motor brain allows precise vocal communication?
S. teguina
Small highly vocal tropical rodent native to cloud forests of Central America. Related to Peromyscus genus and other New World rodents. Of the Crecetidae family (voles, hamsters) and of the same superfamily as house mice and Norwegian rat (Muroidae).
Male and female produce vocal sequences consisting of a series of discrete, frequency-modulated elements strung together, which characteristics that change predictably as the vocalisation progresses.
Song and notes
Each vocal episode is a song; the individual components of the vocal episode are considered notes.
What question(s) is the figure trying to answer?
–> What characterizes S Teguina’s vocal sequences?
Methods:
–> Recorded vocal interactions and created spectrograms and trajectories
1A: S. teguina in its natural habitat.
1B: Spectrograms; frequency (Hz) as a function of time. The three boxes are spectrograms of different sample notes from one animal. Intensity/amplitude is denoted by the colour (blue = lower amplitude; red = greater amplitude).
1C: Spectrogram of a full song; the coloured arrows show the insert time of the corresponding notes from B.
1D: song trajectory plot; plotted the duration of each note as a function of its onset time within the song. Coloured circles represent the notes from B. Song trajectory plot provides a succinct representation of motor sequence.
Findings: (combined with 1E)
–> Males and females produce vocal sequences consisting of discrete frequency-modulated elements strung together, with characteristics that change predictably as vocalisation progresses
What question(s) is the figure trying to answer?
–> Are the acoustic characteristics of S teguina modulated by social context, like we see in other taxa? (Fig 1E, F, G)
What methods are used in the experiments in the figure?
–> Staged a social encounter: male subject (recruit) went into room with male that had been there for 1+ weeks (resident). Mice were in adjacent rooms, could hear but not see each other.
–> Then measured different characteristics of the vocalisations, like note durations, not onset times (taking song trajectory plots), and measuring the number of songs/hour produced.
1E: Song trajectory plots from one male (recruit?) mouse in different social contexts (15 songs per condition). When mouse was alone, the trajectory plots of its vocalisations were highly stereotypes, so not much variation in the vocalisation sequences when alone. The vocalisations are highly predictable. We see this vocal stereotypy in both alone conditions
- Song trajectory plot contains individual dots (which represent the duration of each displayed song) and histogram (which quantifies the duration of all songs produced in a given social context - very similar in alone conditions, much almost 4x more in social condition (57 vs 388 vs 50 songs). Red line is song trajectory from 1D.
–> Variability of song trajectory plots increased significantly when recruits could hear resident mouse (more dots, which are more spread out = variability is song durations; and trajectories show differences within the song in terms of both not onset times and also note durations).
1F: Comparing number of songs per hour for the 8 recruit mice in the different social conditions. Can see statistically significant increase in the number of songs per hour in the social vs alone conditions (4x as often in social vs alone).
–> number of songs per hour (y-axis) defined as SD of song duration distributions
–> Degree of social engagement measured by countersinging and variability
1G: Song duration variability was higher in the social context than in isolation. Statistically significant difference between the alone conditions and the social condition.
–> red line is example mouse from E.
Findings: Social context modulates vocalisations in S. teguina.
Males and females produce vocal sequences consisting of discrete frequency-modulated elements strung together, with characteristics that change predictably as vocalisation progresses (Fig 1 A-D)
Recruits vocalized 4x as often in social context compared to isolation (Fig 1 E, F)
Variability of song trajectory plots increased significantly when recruits could hear resident mouse (1E)
Song duration variability higher in social context than isolation (1G)
What question(s) is the figure trying to answer?
- What is the fine structure of vocal interactions between male mice?
What methods are used in the experiments in the figure?
–> Simultaneously recorded songs of both resident and recruit in social condition
–> Aligned interaction bouts to songs of resident → found recruit mouse times vocal onset to coincide with end of resident’s song (2B) –> Probability distributions sharper at end, suggesting recruit uses end of song
–> To estimate amount of countersigning, they shuffled the song times and quantified likelihood of “spurious countersinging” to an order of magnitude less
2A:
Shows us the interaction setup and the audio recordings of one hour of continuous recordings. You have a “recruit” and “resident”. Recording exclusively in the social condition. Exchanges can be initiated by either male, although the exchanges are typically ended by the recruit.
2B:
Vocal interactions from pair over 24 hr period, with corresponding start and stop probability distributions of recruit. The vocal interactions are aligned either with the end of the resident’s song (left) or beginning of the resident’s song (right). There were 101 vocal interactions.
We can see that there isn’t too much overlap between the black bars (resident’s vocalisation) and the red bars (recruit’s vocalisations).
–> We can see that exchanges usually end with recruit’s song even if initiated by either mouse –> asymmetry seen in all pairs and preserved throughout the 24 hour session.
–> Alignment allows us to see how precisely the recruit mouse is timing his vocalisation with the end of the resident’s.
–> Probability of the recruit singing vs. the timing of the resident’s song offset or onset. We see that the probability of the recruit singing before the resident finishes is very low, and jumps up right after the resident’s song offset (very high immediately after offset, then drops off). The probability of the recruit’s song is high right as the resident starts his song, and drops alongside the onset of the resident song before rising up again as the resident song progresses and comes to an end.
–> The recruit’s response prob- ability distributions were significantly sharper when interaction bouts were aligned to the end of the resident’s songs rather than the start, suggesting that the recruit mouse uses the end of the resident’s song as a sensory trigger
2C/2D:
Summary of mean start and stop latencies across all pairs w.r.t offset of resident’s song.
–> Shows us how findings in B are consistent across all 8 pairs. On average, recruit picks up song right after resident stops; and recruit stops (generally) right after the resident starts.
–> Horizontal line is a song jitter.
What question(s) is the figure trying to answer?
- What is the fine structure of vocal interactions between male mice? What is the/is there temporal coordination of vocal interactions between male pairs?
2E:
Probability of song occurrence aligned to end of resident mouse’s song for pairs (top, example pair in A; bottom, all pairs).
–> Recruit males capable of actively timing vocalization onsets and offsets to avoid resident vocal overlap (turn-taking dynamics). We can see that there is rarely overlap –> activate avoidance of song overlap by the pairs.
–> the end of the song is way more influential in the timing of the recruits song. There’s more variability during the start of the song because both resident and recruit can start the song
2F:
Song initiation jitter is negatively correlated with countersinging probability. Jitter = lack of precision in song. When there’s more social engagement (countersinging probability), the recruits are more precise(less jittery) in their countersinging.
–> Each dot = behaviour of one recruit mouse.
2G:
Song initiation jitter is negatively correlated with the degree of song duration variability change from the alone condition. When the recruit’s song has more variability in duration (social engagement), the recruits are more precise (less jittery) with their countersinging.
Findings:
Found extensive temporal coordination of singing behavior within vocal pairs.
Either recruit or resident male could start exchange, but recruit typically ended (Fig 2 A-E).
Preserved for entire 24hr session (B & E).
Context can influence the timing of vocal turn-taking in other species.
SO:
–> Social context influences the precision/timing of countersinging and characteristics of a mouse’s song. There is temporal coordination of vocal interactions in male mice pairs.
What question(s) is the figure trying to answer?
–> What are the neural mechanisms of countersinging? (3A-C)
What methods are used in the experiments in the figure?
–> Used electromyography to characterize biomechanics of song production (looked at motor movement). EMG helped look at relationship between brain centers and song-related musculature.
–> Used intracortical microstimulation (ICMS) over anterior cortex to identify areas that result in flexion of song-related muscles.
3A:
Electromyograph from muscle and raw audio of an advertisement song.
–> Increased muscle activity immediately before vocalization of individual notes
–> Flexion of jaw muscle (digastricus) before individual vocalizations during exhalation
–> So, there’s a correlation between song production and jaw movement. Means that they can now determine which areas might innervate/control song-related musculature.
3B:
Administered ICMS over large portion of the anterior cortex; minimum current that reliably elicited a fixed EMG activity (so eliciting activity in the jaw musculature related to song production) was used to define the functional hotspots mapped onto the anterolateral aspect of the motor cortex (3C).
Showing ICMS of 2 diff loci → elicits short-latency EMG activity.
–> The parallel lines are represent the stimulation (at different currents - 90 uA vs 10uA vs 30uA), and the activity to the right is EMG activity from the jaw musculature.
3C:
Motor cortex. On the left, showing different ICMS sites, colour-coded according to the threshold current (the minimum current that reliably excited a fixed EMG activity threshold). Left shows example mouse, right shows population data. Population data shows that there’s a hotspot on the anterolateral motor cortex. This region corresponds to the orofacial motor cortex is Mts musculus = OMC.
Overall = probe different areas within the motor cortex with ICMS and administer different voltages, and see at what threshold of voltage they can elicit EMG activity; the lower the threshold the higher the sensitivity. And in some areas no amount of voltage could elicit EMG activity, which means those areas are not responsible for/do not innervate the jaw musculature.
What question(s) is the figure trying to answer?
–> What is the functional role of the OMC on song production? (3D, E, F, G)
–> while they can see that OMC stimulation elicits electrical activity in song-related musculature, they don’t know whether the OMC can actually influence song production itself.
What methods are used in the experiments in the figure?
–> Perturbations during singing alone (bilateral electrical stimulation of OMC) (3D)
–> –> For every trial, used 10 notes before perturbation to estimate note durations that would be expected in an uninterrupted song.
Then compared actual note durations with predicted values (3F).
3D:
Example spectrograms from one mouse while perturbation was administered during its song. In different trials with different strength of electrical stimulation to the OMC, song was truncated or paused.
In the strong stimulation, song was truncated (we see that frequency is 0 after stimulation period).
Milder stimulation (200 uA -500uA) often produced brief pauses, with songs resuming after stimulation ended.
SO –> OMC influence song production, since perturbation messes with song production (halting or pausing it).
3E:
Since we know that S teguina songs are precisely stereotyped, we can distinguish the effects on song production caused by OMC perturbation and therefore suggest the role played by the pathway from the OMC to vocal musculature.
–> Illustrates two possible outcomes for song resumption after brief pause caused by perturbation:
Model 1 (red) : If song resumes at expected point in sequence (accounts for time delay), suggesting vocal patterning is primarily driven by a pathway independent of OMC (Do you ever feel like a plastic bag | PERTURBATION |…wanting to start again).
Model 2 (green): song could resume at same point in motor sequence where it paused, suggesting pathway leading from OMC to vocal musculature sculpts song structure. (Do you ever feel like a plastic bag/ | PERTURBATION/| drifting through the wind …).
3F:
Full trajectory of an example song before and after electrical stimulation-induced pause.
–> Dots indicate the duration of each note.
Zoom in on the box shows what we would expect the plot to look like based on the two models. Grey line is actual change in note duration. The actual plot fits the trajectory of model 2: the song picks up where the mouse left off before the perturbation.
3G:
- Summary data for all paused songs in each animal. Red line shows model 1 expected trajectory, green line model 2. Most trajectories are consistent with model 2.
SO: refutes the hypothesis that song patterning in S teguina is shaped by an OMC-independent pathway. (the pathway must be OMC-dependent in some way).
OVERALL: OMC does in fact have a functional role in song production, in that it appears to have a role in song patterning, because perturbing it causes the song to stop and it only resumes its progression when perturbation stops.
What question(s) is the figure trying to answer?
–> What is the nature of the interaction between the OMC to behavioral output? (3H - K)
What methods are used in the experiments in the figure?
–> Used mild focal cooling of OMC during sound production to slow OMC dynamics and isolate contributions of local neurons in OMC (3H-K). Isolate local neuronal dynamics from downstream. (Maintains behaviourally relevant activity while slowing these dynamics relatively).
3H:
3 models/proposed effects of OMC cooling on song trajectory if OMC activity doesn’t affect song timing, if it exclusively controls song timing, or if OMC and subcortical structures share this control (so 3 models of the pathway involving the OMC and vocal musculature).
Model 1: If OMC doesn’t affect song timing, control and cooled song trajectories should completely overlap.
Model 2: If OMC dynamics exclusively affect song timing, cooling should lead to dilation of vocal behaviour on all time scales (note duration and song length). - slowing OMC dynamics should slow the vocal behaviours.
Model 3: If motor control is shared with other subcortical regions, cooling should alter only some temporal properties.
3I:
Spectrograms and trajectories of example songs during baseline and cooling sessions. Cooling of the OMC lengthens song durations by decreasing the rate of change of note duration (slope) during songs.
–> OMC cooling resulted in shallower song trajectory that took longer to unfold.
–> What we see in the spectrograms is that the songs duration is longer.
–> Song trajectory shows us that the slope is shallower - the rate at which note duration changes throughout the song is lower than normal (the more cooling is applied).
3J/K:
Cooling of OMC lengthens song durations by decreasing rate of change of note duration (slope) during song. Cooling also resulted in an increase in the the needed to reach a threshold note of 75 ms without changing the duration of individual notes (so progression through the song overall took longer).
–> Arrows represent mean values of each distribution.
–> K shows population data, each of the 8 mice.
Since the distribution of individual note durations did not change, OMC does not solely determine all aspects of song timing.
–> OMC shapes song progression without influencing the structure of individual notes; it takes longer for the progression of notes to occur, which is done via increasing the total number of notes produced.
OVERALL:
Suggests that there is a hierarchy of motor timing control (model 3 of 3H). The OMC exerts moment-by-moment control over the pacing of a sub-cortically generated song sequences.
What question(s) is the figure trying to answer?
–> Does the OMC contribute to social coordination by altering song structure? (They know that social interaction changes song progression and song variability is driven by degree of social engagement, and that the OMC is well positioned to contribute to that social coordination since it exerts control over the pacing of the songs).
(Social coordination refers to social interaction changing song progression (1E) and degree of social engagement during vocal interactions driving song variability (2F, G)).
–> Does OMC mediate social influences on singing behavior?
What methods are used in the experiments in the figure?
–>Tested first question by reversibly inhibiting OMC with muscimol (GABA agonist)
–> Muscimol = OMC inactivation
–> Lowered concentration of injections to make sure that nonspecific motor deficits were not evident.
–> each animal got both experimental and control conditions (saline).
–> Tested second question by using playback
4A:
Set up of social simulation/influences. Had song playback to mimic a resident and recorded the mouse’s response. Diagram shows example countersigning response to audio playback.
4B:
Design of experimental paradigm.
Animal is injected with either saline or muscimol, then left for a pre-testing period, where they measured spontaneous singing (singing without prompt by “resident” or audio playback). Then playback starts, and they measure singing during the testing period (1 hour).
4C:
Song rastor plot of all trials with either saline or muscimol. Each row is a different session. Ticks are playback from speaker, dots are mouse songs. Bar plots on the left = total number of spontaneous songs during pretesting period (nonsocial).
5/6 mice produced spontaneous songs, and rate of spontaneous singing was not significantly influenced by this manipulation (meaning injecting muscimol?).
In control animals (red), song playback led to increased amount of singing as well as song duration variability (expected in countersinging).
Muscimol-injected mice did not sing more songs in response to playback.
–> suggests that OMC affects context-dependent modulation of song rate. (common in natural social encounters — 1F)
4D/E:
Compared with saline (control) injections, dosing with muscimol eliminates a precise countersigning response.
Showing the bouts of signing and the alignment, also showing probability of songs as function of the time - playback song ends. The probability of eliciting a countersigning response was significantly greater in the control condition than in the OMC-inactivated condition for each mouse and across the population. And if there was countersinging, there was an increase in mean response latency.
4F:
Countersigning response probability significantly decreases with muscimol treatment compared with saline dosing.
4G:
In control animals, the total number of songs per house significantly increases during playback period. This increase is absent when OMC is inactivated. Gary circles = individual animals. Black circles = mean and SEM.
4H: Song durations during the playback period are significantly higher for saline dosed mice compared with muscimol injected.
SO WHAT:
OMC is critical for rapid vocal responses to playback; such responses must be driven by sensorimotor coupling rather than by more general changes in motivation.
