Brain imaging reveals how learning process drives learning itself

Dec 6, 2005

The term "judgments of learning" (JOL) describes a neurological process that guides us as we turn new information into old information. JOL helps us learn information in the short term and determine whether we'll remember it in the long term. In a functional MRI (fMRI) study, psychologists evaluated the neural basis of JOL, specifically looking at brain circuitry.

"The ability to successfully encode experiences into long-term memory has been associated with the functioning of medial temporal lobe (MTL) and prefrontal cortex (PFC)," wrote Yun-Ching Kao, Ph.D., and colleagues from the department of psychology at Stanford University in Stanford, CA, and the Massachusetts Institute of Technology in Cambridge. "MTL structures are essential for the formation of new declarative memories.... Studies have implicated PFC as important in JOL accuracy (but) it is unknown which specific subregions support JOL" (Nature Neuroscience, November 13, 2005).

The present study included six female and 14 male participants. They were asked to look at 700 pictures of indoor and outdoor scenes, both during and post MRI. After analyzing a picture, an individual had to predict whether he or she would remember each scene later or forget it.

MR scanning was done on a 3-tesla unit (Signa, GE Healthcare, Chalfont St. Giles, U.K.). Anatomical images were acquired first using a spin-echo T1-weighted protocol. Functional images were obtained in the same location using a T2^*-sensitive, 2D gradient-echo spiral in/out sequence. Each scanning session lasted nine minutes.

This experiment had four possible outcomes:

A "will remember" prediction that was actually remembered
A "will remember" prediction that was forgotten
A "will forget" prediction that was remembered
A "will forget" prediction that was actually forgotten

On average, the participants made 144 "will remember" predictions and 104 "will forget" predictions. For remembered scenes, activation on MRI was greater in bilateral MTL regions, such as the posterior parahippocampal and fusiform gyri. In terms of an individual's predictions, significant activation was seen in several regions, including the left lateral PFC and left amygdala. The MTL and the left posterior cingulated showed greater activation for actual encoding versus predicted encoding.

The study identified four patterns of brain behavior in regions known for their importance in memory, thought, and introspection.

First, MTL regions did not seem to be involved in encoding success when it came to making "will remember/will forget" predictions. Second, greater deactivations in the medial PFC were associated with subsequently remembered items, implying that once an individual had grasped a concept, he or she would relocate neurological resources to the next task. Third, activation in the ventromedial PFC may have supported JOL predictions by forming an internal model of successful learning. Finally, while lateral PFC activation occurred for successful memory encoding, it did not correlate with JOL accuracy.

"A fundamental issue in human memory research is the relationship between objective and subjective dimensions of memory," the group stated. "Such subjective processes during learning are of particular interest, because these processes can actually enhance the effectiveness of learning... These processes, and the neural circuits that mediate them, constitute a critical component of the way in which knowing how to learn empowers learning itself."

By Shalmali Pal
AuntMinnie.com staff writer
December 7, 2005

Too much information: fMRI indicates lack of discrimination at fault for senior forgetfulness, September 19, 2005