Martha Bosma - The reason for the rhythm: mechanisms for driving spontaneous activity in the developing mouse hindbrain
Today's field trip was to a biology department seminar at the UW. Martha Bosma presented some rather striking work on spontaneous neural activity in the mouse hindbrain; specifically an instance of it during fœtal development.
The general phenomenon that her lab studies is the presence of co-ordinated neural activity without any input to stimulate it. That there is some kind of neural activity without stimulation is neither surprising nor informative—individual neurons have a 'resting rate' of firing spontaneously, so of course a population of neurons also would—but for the firing to be co-ordinated demands explanation. Even if it is purely epiphenomenal, it has the potential to be a useful way to explore the interconnections in the brain, and the way that network structure influences processing.
At least half of the talk was taken up with technical information about how the neural recordings were made, and exactly what the patterns they have observed look like. I'll skip that, partly because I don't think the detail is essential to understand the main message that I took from the talk, but mainly because it's far enough from my area that I fear mangling the information in re-telling. I will say one thing though: every time I hear from someone who experiments with 'wetware' as opposed to sitting behind a desk and running simulations, I am filled with respect for them and their technicians, because the basic mechanics of gathering data are so much more difficult for the path they have chosen than they are for us simulation people.
The specific pattern that was the subject of this talk is observed in the hindbrain of mice, during a very short window of development. The hindbrain is bilaterally symmetrical, and this pattern of activity takes the form of a wave that starts somewhere along the midline and spreads both caudally and outwards, reaching motor neurons that innervate the musculature of the face. Bosma outlined a developmental timetable as follows (numbers are days since conception; mice gestate for 18 days):
- 8.5
- Neurons appear [these may not be the first neurons, but this is when the neurons being studied appear]
The mechanisms of the pattern's disappearance remain an open question; the neurons remain richly interconnected, and continue to respond to all the same hormones. What I found interesting, though, was the postulated significance of a pattern like this appearing for such a short time and disappearing well before birth.
The key issue is that this is a motor control region, and as such there is a systematic relationship between the location of a neuron and the location of the muscles it stimulates. I'm not sure the relationship is as straightforward as the standard motor homunculus diagram suggests, but position is certainly meaningful. The problem is that neurons have to extend connections to the right point, which means that they need to "know" their own location. It's not as if neurons can have a conversation among themselves, so somehow this information needs to be conveyed by purely local signalling. This is where spontaneous rhythmic activity comes in: spreading activation in a fixed direction gives each neuron position information relative to its neighbours.
I don't think the question of why the pattern subsequently disappears was addressed, but the answer to that seems reasonably easy to infer. Once the neurons have grown their dendrites to the appropriate muscle cells, the rhythm is no longer useful, and it risks causing spontaneous twitching in the muscles that these neurons innervate, as well as drowning out meaningful activity in that region of cortex.
For me personally, this talk had some additional value as a reminder of one of the ways my experiments are completely unrealistic. In every variety of agent I've looked at to date, the interconnections of the 'nervous system' have been pre-specified, as if an individual had an adult brain from birth. For the work I'm doing right now this makes sense, because the environment only changes within an individual's lifetime, not between generations as it clearly does in the real world. However, I am ultimately interested in how different time-courses of adaption interact with each other, and development is hugely important in adapting to an unpredictable world.
Comments