When performing, musicians often have to play common motifs in a piece at various tempos. Though understanding of how our brain can produce the fine movements needed to play an instrument is well-developed, the ability to execute these over a wide range of controlled speeds remains unexplained. Recording from the frontal lobe and two connected midbrain regions of non-human primates, the lab of Mehrdad Jazayeri at MIT has recently found a novel - and surprisingly simple - mechanism that could elucidate such speed control.

Rhesus macaques were trained to respond to a cue either by pressing a button or with a saccade to a target on a screen after either a long or short delay. In each trial, a symbol would specify whether the monkey was required to pause for 800 or 1500 milliseconds before producing the appropriate response. During the task, co-first author Jing Wang recorded neuronal activity from the three brain areas previously implicated in both the timing and initiation of movement; the clock activity of these regions was confirmed via an infusion of a drug able to inactivate the relevant cortex of the frontal lobe, and hence significantly decrease the ability of the macaque to pause for an accurate timeframe.

“A key feature of the recordings made by the researchers revealed that individual neurons within circuits could stretch or shorten their activity”

In order to produce a behaviour, an entire circuit must act in concert, with different cells firing at different points. Changing the time at which any subset of neurons is firing therefore affects the total time taken for the circuit to progress through the complete repertoire of activity patterns necessary for either a saccade or button press to be produced. Interestingly, however, the recordings made by the researchers revealed that individual neurons within circuits could stretch or shorten their activity. This is in contrast to the foremost theory until now, that nervous timing relies on a central accumulator which can impose a varying threshold on the neuronal activity required to initiate an action.

“The underlying timekeepers which adjust the speeds of these circuits remain a mystery”

To further investigate these network dynamics, co-first author Devika Narain generated a model of the brain networks studied. This computer simulation was trained on a similar task to that presented to the monkey, and was able to learn to produce the same output as the in vivo circuits. Upon analysis, the components of this simulation shared some key features with those recorded from neurons, and highlighted the potential importance of non-linearities for the neuron’s ability to compress their activity. That is, the model predicted that the ability of neurons to produce an output that is not directly related to the intensity of their input was key to this ability, a finding borne out in the brain recordings.

The work builds on a growing suggestion that the components of specific circuits, and not some general clock, are the timepieces of the brain. Such activity is crucial not just in timing movements, but also for how our brains are able to maintain a reasonable speed-accuracy trade-off when perceiving our environment. However, the underlying timekeepers which adjust the speeds of these circuits remain a mystery, and the focus of future work at the lab-bench