Skip to main content

Neural circuits for optimal prediction of timing

Periodic Reporting for period 1 - PredOpt (Neural circuits for optimal prediction of timing)

Reporting period: 2018-04-01 to 2020-03-31

General problem:
When we observe the world, we do not merely act upon our observations. Rather, we use our prior experience with the environment to make decisions that are statistically optimal in most cases. Nowhere is this phenomenon more true than it is for timing. We have strong prior expectations of the speed of cars on the highway, or of the timing of a ball falling under the influence of gravity, or the time a tennis serve will reach our racket. These anecdotal ideas have been studied by decades of psychophysical experiments and formalized for human timing behaviors through mathematical modeling. However, until now, we do not understand how the brain stores and processes these sophisticated behaviors. The goal of this project, PredOpt, was to investigate neural mechanisms of such behaviors that involve rapid timing movements.

Benefits for society:
Cognitive psychology has studied the interplay of memory and sensation in timing for many decades and has developed a good understanding of these behaviors. However, in the clinic, loss of cognition related to temporal skills is complex and it is virtually impossible to concretely diagnose symptoms arising from loss of such functions. This is why in PredOpt we aimed to make steps towards uncovering the neural circuitry behind cognitive aspects of timing, where the memory of temporal experiences is stored and acquired. We aimed to develop mathematical models of these neural pathways so that we could test loss of function and make predictions for what consequences may arise. This has two benefits. 1) It helps us understand how the healthy brain utilizes memory and cognition to perform skillful and well-timed movements. 2) By understanding neural mechanisms that underlie rapid anticipation and prediction during well-timed movements, clinicians can establish better links between symptoms and potential pathways that are causing specific loss of function.

Overall objections:

We had the following overall objectives for PredOpt:
1) To identify a neural pathway where the memory of sub-second time intervals is acquired, stored and processed.
2) To manipulate the experience of the animal and test whether the pathway reflects this change in experience
3) To develop a mathematical model of the neural mechanism by which this experience is acquired and generate predictions for future clinical and neuroscientific investigations.

This work is performed in the mouse model because the expected pathways responsible for these behaviors lie in the cerebellum, which is a structure that is phylogenetically preserved in all mammals and therefore we expect a basis for clinical translation to primate models.
Overview of results achieved so far:

1. We have identified a pathway in the cerebellar cortex where sub-second time intervals are acquired. Using a cutting-edge technique called optogenetics, we can activate or silence cells (neurons) responsible for encoding memories of these time intervals and contrast this with behavioral symptoms of the animal showing a clear loss of function. We have performed control experiments to suggest that this loss of memory is specific to this area.

2. We recorded the electrical activity of neurons in the identified area while we altered the temporal experience of the animal. During this time, we observed a change in behavior and the formation of a new memory and simultaneous adaptation of neural activity in response to this shift. These results further demonstrate that this neural pathway is responsible for the acquisition, modification and usage of sub-second time interval memory.

3. We have developed a mathematical model that fully predicts the observed findings and makes predictions of a neural mechanism for how this part of the brain acquires temporal information and converts it to memory. We call this model TRACE 2.0 since it revises an earlier mathematical model, developed by the candidate.
Scientific advance:
1. This is the first instance, in our knowledge, that has demonstrated a neural mechanism for the acquisition of prior distributions of sub-second time intervals and their memory.
2. We have discovered a new neural signal during our experiments that sheds new light on how the brain learns to encode new time intervals and form new memories.
3. We have developed a new mechanistic model that explains our existing findings and makes predictions for future investigations.

Societal impact:
1. Having developed a model of a mechanism for timing behaviors in the sub-second range, we have reached out to clinicians to start a conversation about our model's predictions for behavioral symptoms when different parts of the pathway are disrupted. The models makes several clinically relevant predictions about loss of timing functions.
2. Our findings entail new discoveries and our model makes predictions for future neuroscientific experiments that we hope the field will engage with and utilize for future investigations.