Differential Encoding of Time by Prefrontal and Striatal Network Dynamics

Konstantin I. Bakhurin, Vishwa Goudar, Justin L. Shobe, Leslie D. Claar, Dean V. Buonomano, and Sotiris C. Masmanidis

The Journal of Neuroscience （2017）

DOI:10.1523/JNEUROSCI.1789-16.2016

The conceptual gap or the scientific question: how striatum and orbitofrontal cortex (OFC) encode time?

How the authors proposed the question: I guess this was a "post hoc" story, that is, the authors did the experiments first (with one idea held in mind, maybe about working memory), but wove a new story at last ("time encoding").?

Brief summary: by training mice to learn a trace Pavlovian conditioning task, the authors shown, with extracellular recording,?that the medial spiny neurons in the striatum could encode task timing, better than that of the OFC neurons, in a lick timing dependent manner.

How the authors tell the story: one major point the authors tried to sell to the audiences is that no previous study had compared the coding ability of timing between distinct brain areas within the same behavioral context. But in my point of view, it's not a good selling point. It's just a small step of advance of recording technique.

The most serious shortcoming of this study is that the task they used is not a real "timing" task. Because there are no manipulation of the timing interval between the stimulus and the reward, they can not definitely tell the timing is a critical parameter for mice to perform this task, especially by considering the unsynchronized licking onset time. If they can successfully trained mice to lick at distinct time point to get reward with distinct conditioned cues, that will be a much more convincing "timing" task.?

Under the current task design, we can say that the so-called "time encoding" is a confounding observation of the licking response.

Although I do not buy the conclusion, but I like this idea. To me, it's a totally new framework for framing a story. I learn something new from it.

The ongoing question(s) inspired by the current study: inspired by this study, I conceive a better behavioral design for studying "time encoding": is it possible to train mice to lick at 1 or 3 or 5 or 7 seconds after 4 distinct conditioned cues? We can set the response window half a second before and after those time points for measuring the correctness of the behavioral response. According to my experience, it is doable. Maybe I can test it with my own mice. I like this idea.