The motivation for this model is to provide a starting point for the search for the "musical genes". I want to know what is the simplest mechanism which can result in the perception of music.
Single-Protein versus Multi-Protein or Zero-Protein
The model to be described here is "single-protein" in the sense that:
- There is one human-specific protein which is responsible for detecting the "musicality" of music and passing that information on to parts of the brain which respond to that musicality.
- This protein may be a variation of an existing protein, which, in its original form, performed another task not particularly related to music perception, but which when altered, does perform that task as a side-effect of the variation in how it performs its primary task.
It is quite possible that one protein is not enough, and that the perception of music involves at least two proteins. However, from an evolutionary point of view, there must have been one protein which was first to evolve in the direction of human musical perception.
In general I think it is useful to consider the simplest possible model that might explain something, just for the sake of exploring how simple the explanation could be, and perhaps to discover by some means that some slightly higher level of complexity is indeed required.
An even smaller number than one is zero, and it is possible to imagine a zero-protein model of music perception, where music perception is entirely determined by non-proteing encoding genes which, for example, determine the expression of existing proteins in a manner which controls the connectivity between different types of neurons resulting in the perception of music.
The model I am presenting assumes the following informational "circuit":
- Perception of individual aspects of musics results in "constant activity patterns" which are perceptibly constant over a time-scale in the range of one to ten seconds.
- These constant activity patterns can be observed by observing the contrast between the inactive portions and the active portions of the pattern.
- Neural activity is correlated with the concentrations of various molecular species in the extra cellular medium, with concentrations being correlated over different temporal and spatial scales.
- For the purpose of the model, assume the existence of two such molecular species, A and B.
- The concentration of molecule A is correlated with neural activity with spatial resolution similar to that of astrocyte activity, and with the required temporal resolution.
- The concentration of molecule B is similar, but it has a reduced spatial resolution, perhaps because it diffuses more rapidly through the extra-cellular medium.
- In effect the concentration of A represents activity over a very local area, and the concentration of B represents average activity over a somewhat larger area (perhaps 2 or 3 times larger).
- The contrast between the relative concentrations of A and B provides information about the contrast between neural activity in the current location and neural activity a small distance away.
- The special protein (which we can call Protein M, where "M" is for "musical"), is located on the synapses of cortical neurons which project onto the amygdala, and it plays some role in the synaptic response to some particular excitatory neurotransmitter (e.g. glutamate, but it could be some other neurotransmitter).
- Therefore, when constant activity patterns are detected, protein M increases the synaptic response of the neuron it is on, which in turn activates some part of the amygdala.
- The consequential increased activity of the amygdala influences the consolidation of long-term memory, and in particular it modulates the operation of uncritical listening, i.e. the long-term accumulation of a world-view absorbed by the mechanism of believing (to some extent) that what other people say is true.
This particular model implies certain constraints on the participants:
- The input (afferent) synapses of the amygdala-projecting neurons must be co-located with the cortical maps where the constant activity patterns occur. (This is mostly in the auditory cortex, but the super-stimulus theory can apply to non-auditory aspects of speech perception, such as the perception of gesture and body language, which may account for some aspects of human response to dancing.)
- Protein M must be in a position to modulate response to neurotransmission across the synapse, but it must also be exposed to the extra-cellular medium.
- Protein M must be a human-specific variation of some non-human protein, i.e. M-precursor, which itself plays an existing role in the modulation of response to synaptic signalling.
- The modulation of synaptic response by protein M must be on a time scale not significantly slower than the 1 to 10 second time-scale over which constant activity patterns are to be perceived. (It can be faster, but being much faster will make little difference because the overall response time-scale is limited by how quickly the concentrations of molecules A and B can change in response to neural activity levels.)
Candidates for M-precursor, i.e. for types of protein that protein M could be a variant of, include:
- Synaptic receptor proteins
- Proteins which for ion channels
- Neurotransmitter transporters
- Neurotransmitter transporter regulators
Although the model above specifies action on excitatory synapses, it is also possible that protein M acts on inhibitory synapses, in which case it would have to act in the opposite direction, i.e. contrast between relative concentrations of A and B would result in a reduction of inhibitory response in the synapse.
Is there a "musical neurotransmitter"?
The single-protein model is in effect a zero-neurotransmitter model. This is because although protein M is acted on by molecules A and B, it responds only to their relative concentrations at one particular location, and there is no sense in which the relevant information can be transmitted from further away.
A Two-Protein Model
If we consider the possibility of some kind of musical neurotransmitter, then the model becomes a two-protein model, i.e. one protein M1 to release the neurotransmitter, and one protein M2 to receive it.
In this extended model, protein M1 is the one that responds to the relative concentrations of molecules A and B, and consequentially increases the amount of the musical neurotransmitter MN in the extra-cellular medium. Protein M2 is situated on the afferent synapses of the amygdala-projecting cortical neurons, and it contains a domain which responds to the presence of MN in a manner which modulates its normal function.
In principle the transmission of MN might occur over an arbitrarily large distance across the brain. However in practice the transmission may be limited to a relatively short distance, for at least two reasons.
The first reason is that transmission of MN is entirely by diffusion, and there is still a requirement for the relevant information to be transmitted within a required timescale of seconds at the most.
A second reason is that at least one of the proteins M1 or M2 is likely to be more "music-specific" than the other. Again, considering the evolutionary origin of music, one of the two proteins had to evolve first, as a variant of some other protein, i.e. either M1-precursor or M2-precursor.
One possible scenario is the following:
- Protein M1-precursor exists in astrocytes where it responds to relative concentrations of A and B by releasing MN.
- The molecule MN serves some pre-existing signalling purpose. Given the general local nature of the "responsibilities" of astrocytes, in maintaining and perhaps co-ordinating the activities of co-located neurons, this signal is likely to only have local significance.
- Protein M2-precursor evolves into protein M2 which has a domain which responds to the presence of local MN, altering the normal operation of M2.
Possible Types of Protein Variant
Both the one-protein and two-protein models require the evolution some human-specific variant of a pre-existing brain protein.
Various forms of protein variant are known, including:
- Gene duplication, where the gene for protein M-precursor would duplicate into two copies, and one of them evolves into a gene for protein M.
- Retrogene formation, a special form of duplication, where a retrovirus recreates a DNA gene from an mRNA transcript.
- Alternative splicing, where the "same" gene is transcribed differently in different circumstances, for example including or not including certain introns, or selectively applying other transcript editing operations.
It is a surprising fact about the human genome that very few human-specific brain-expressed protein genes have been discovered.
The human genome has been fully sequenced, and protein genes can usually be directly identified according to:
- Homology with other genes, either human or non-human
- Promoter regions
- Evolutionary conservation
It follows that most obvious duplicate protein brain genes would already have been found, unless they are very cleverly "hiding" from scientific view.
However alternate splicing is one form of duplication which is less obviously detectable, involving as it does duplication of a gene in its current location, and this remains a possibility for discovering as-yet undiscovered human-specific brain genes.
One example of a human-specific variant protein due to alternative splicing is Neuropsin II (the nature and expression of this variant protein rule it out as a possible candidate for any of the M proteins in the models just described).
One apparent example of a fully duplicated human-specific brain-expressed gene is SIGLEC11 (and see also the OMIM page), but this seems to be more relevant to immunity than to information processing, as it is expressed on the microglia which are part of the immune system.
Various forms of human-specific non-protein "genes" have been discovered. For example the paper Accelerated Evolution of Conserved Noncoding Sequences in Humans describes the discovery of human-specific promoter sequences which promote the expression of neural-adhesion proteins. Such genes are likely to be involved in the human-specific patterns of wiring in the brain, and they demonstrate how humans can have evolved in very human-specific ways without evolving any new actual proteins.
Unfortunately this kind of model does not help to explain music, at least not within the framework of the super-stimulus theory, which requires the "observation" of patterns of neural activity. This type of observation is most directly achieved by "observing" concentrations of activity-related molecules, which is the sort of thing that only a specialised protein can do.
One could perhaps imagine the evolution of some novel neural connectivity patterns which would support the observation of constant activity patterns, however one would expect this to produce some anatomically obvious difference between human and non-human brains, especially in the auditory cortex, and no such difference has ever been observed, to the best of my knowledge.