Monash

Discovery

When predictions go wrong

22.12.2015

In a nutshell: Identifies a new brain circuit that detects the unexpected.

View Paper Abstract
When predictions go wrong

The big picture:

The brain is constantly dealing with extraordinary volumes of incoming data from its senses — by some estimates it faces as much as 11 million bits per second of incoming information. One explanation for this feat is that the brain is an efficient learning and prediction machine. Rather than going to the trouble of processing each bit of incoming data, the clever brain homes in on the differences between its predictions about what those sensory inputs will be, based on past experience and the actual inputs. The still-mysterious process behind this ability is termed “predictive coding.”

Now, a team lead by CIBF chief investigator Marta Garrido at the Centre for Advanced Imaging at The University of Queensland in Brisbane, has discovered a new circuit that makes up part of the proposed predictive-coding machinery.

Garrido’s team used a technique called magnetoencephalography or MEG to record the activity in the brains of people watching a series of pictures that appeared in a predictable order (for example, ABCD), in a so-called mismatched order, where two letters in a routine series are transposed (for example, ABDC), and in an entirely random orders (for example, CADB).

MEG imaging found that in response to mismatch — but not random — orders, both the ventromedial prefrontal cortex (vmPFC) and the hippocampus showed a pattern of slow brain activity called theta wave firing that indicated a circuit between the two brain regions. The circuit was not activated by random or predictable orders.

And what really surprised the team was that analysis of the MEG data suggested that not only is the vmPFC involved in predictive coding, it actually drives the process in the hippocampus, rather than the other way around, as previously thought.

The hippocampus plays a fundamental role in memory, and was already known to play a role in detecting prediction violations. But the vmPFC— a brain region that’s involved in memory, risk and fear, decision making and more — had not previously been identified as a key player in the detection of prediction violations.

The vmPFC could play a role in accessing the memories that the brain needs to make successful predictions, suggests Garrido.

Next steps:
The researchers are investigating how the new circuit’s firing pattern in response to unexpected events changes if our brains are paying attention or focused on other tasks.


Reference:
Garrido, M. I., Barnes, G. R., Kumaran, D., Maguire, E. A., & Dolan, R. J. (2015). Ventromedial prefrontal cortex drives hippocampal theta oscillations induced by mismatch computations. Neuroimage, 120, 362-370.


Republish this article:

We believe in sharing knowledge. We use a Creative Commons Attribution 4.0 International License, which allows unrestricted use of this content, subject only to appropriate attribution. So please use this article as is, or edit it to fit your purposes. Referrals, mentions and links are appreciated.

CIBF