 |
 |
 |
Neural Systems for Memory How does the brain remember where I parked my car? Princeton neuroscientist Kenneth Norman investigates how memory – especially episodic memory -- works by simulating the neural systems that underlie it. The three brain areas agreed to be of special importance in memory are the hippocampus, the posterior cortex, and the prefrontal cortex. Each of these areas is distinct in terms of architecture, neuronal properties and connectivity. Remaining faithful to these biological characteristics, Norman and his colleagues create computer simulations that behave in ways that are analogous to the brain. Most recently, the researchers have been working on a new learning algorithm to address a puzzling phenomenon regarding memory. Studies show that unwanted memory traces that compete with the sought-after memory trace (at retrieval) are punished, in the sense that they become harder to retrieve in the future. For example, if, in the course of trying to retrieve the word "pear," the neural representation of "apple" is briefly activated, "apple" becomes harder to retrieve. To date, few alternative algorithms or explanations exist that satisfactorily capture the counter-intuitive details of these findings. The researchers hope to apply their new algorithm to explain related results in sleep, cognitive dissonance, self-perception and negative priming. On a different track, Norman and colleagues have worked to improve on functional magnetic resonance imaging (fMRI) data analysis. Using a neural network classifier, they’ve shown that it’s possible to discern broadly what kinds of images a person is looking at – a shoe, a house or a face -- just from the pattern of activation in their temporal lobe. This could open a new window into the workings
of areas like the prefrontal cortex, and eventually, how researchers
might represent our current mental context.
|
 |
|
 |
|||
