New Computational Method Helps Decode Complex Circuitry of the Brain

New Computational Method Helps Decode Complex Circuitry of the Brain

Posted: June 10, 2013

Story highlights


A team of neuroscientists, including NARSAD Young Investigator Grantee Melissa R. Warden, Ph.D., of Stanford University, devised new analytical methods to decode neuronal interactions in complex cognitive tasks. Their findings, reported May 30th in Nature, recommend a shift of focus in research studies to analyze neurons in the prefrontal cortex (PFC) that encode all task-relevant aspects before directing a behavioral response to stimuli (‘mixed selectivity neurons’), rather than highly specialized neurons that encode a single aspect. They expect this type of finer analysis will not only enhance understanding of the complex circuitry of the healthy brain, but will also shed light on the circuit dysfunction underlying the cognitive deficits manifested in illnesses such as schizophrenia.

While some neurons in our brain are responsible for integrating external stimuli, or cues, to produce a mechanical response, much of our complex behavior cannot be simply explained by mapping specific cues and their corresponding responses―the same cue can produce different behaviors based on the context. For example, if you carry a cell phone and it rings, it is appropriate to answer it in some contexts, but not in others. The ring itself (the cue) does not automatically determine the appropriate response―our brain usually synthesizes the cue of the external stimulus with other factors in the situational context before responding (e.g., am I driving and at risk of having an accident if I answer the phone? or is this an appropriate time/place to take the call?). 

Many neurons, including the ones in the prefrontal cortex (PFC) of the brain, have complex and diverse responses that can be simultaneously connected to multiple aspects of the task to help us respond in an appropriate way in these situations. This is important, because the PFC receives inputs from all sensory neurons and directly connects them to the final motor response (behavior).

Dr. Warden explains, “One prefrontal neuron might simultaneously encode both the stimulus (cell phone ring) as well as the situational context (don’t pick up while driving). Many single prefrontal neurons encode several aspects of the environment. This property has been tricky for neurophysiologists to understand, since they usually study a neuron’s response to only one kind of stimulus, but it has become increasingly clear that this type of encoding is widespread―especially in higher association cortex like the prefrontal cortex of the human brain.”

Dr. Warden and her colleagues used computational methods to decipher and analyze information obtained from two monkeys that were trained in a cognitive task. The analysis of neuronal activity demonstrated that PFC neurons perform many functions, not just one, and that these mixed selectivity neurons seem to be a hallmark of the PFC and other brain structures involved in cognition. The new analytical methods developed by the authors revealed that the kind of mixed selectivity exhibited by these neurons allows networks in the PFC to encode as much information as would be possible with highly specialized neurons encoding a single aspect, and in fact can confer a significant computational advantage. This development opens up a new direction of research to better understand the network dynamics of the PFC involved in cognition and behavior.