Hierarchical Coding for Sequential Task Events in the Monkey Prefrontal Cortex. Sigala, Kusunoki, Nimmo-Smith, Gaffan, Duncan. PNAS 2008


Like the other papers read today, the notes on this are on the light side, although the paper itself is pretty short.

  1. Examined activity of a set of neurons during a hierarchical task
  2. “For different phases, there were different, approximately orthogonal patterns of activity across the population of neurons… By orthogonal coding, the frontal lobe may control transitions between the discrete steps of a mental program; by correlated coding with each step, similar operations may be applied to different stimulus content.”
  3. “Prefrontal neurons are known to code many kinds of task relevant information (9, 10). They can show selective responses during the different phases of complex tasks, including stimulus
    presentations, delay periods, responses, and feedback intervals (11–16). Within each task phase, there can be selective coding of stimulus and response identity, working memory contents, task
    rules, and rewards (16–20). Even when neurons are randomly sampled, a large proportion code some form of task-relevant information (21). Together, these results suggest a dense, broadly distributed code of information that bears on the animal’s current activity (9, 10).”
  4. So, how is the representation organized in neurons
    1. Orthogonal representations minimize interference
    2. Correlated codes imply similarity, allows for generalization
  5. “The results show that successive task phases are coded by successive, approximately orthogonal activity vectors. Within each phase, activity patterns are strongly correlated for different stimuli, suggesting that stimulus information is coded by modulation of the basic task phase vector. This scheme would allow independent control of the different operations required in successive task events, whereas within each task phase, similar operations are applied to different stimulus identities. Such hierarchical coding may provide a fundamental structure for prefrontal control of sequential activity”
  6. Successive steps in the task aren’t related – there isn’t any smoothness in the task structure
    1. <I would expect that in a task where transitions between phases are smooth, there wouldn’t be an orthogonal representation, as it would hurt generalization and transfer between smoothly transitioning phases>
  7. “The findings show that in each task phase, there was reliable coding of stimulus information. Like overall activity vectors, however, stimulus preferences were rather independent in different task phases. Cue and target preferences, in particular, were almost entirely unrelated, although both were somewhat predictive of delay preference.”
Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: