Endogenous human brain dynamics recover slowly following cognitive effort

Anna Barnes,John Suckling,Ed Bullmore

doi:10.1038/npre.2008.2665

Abstract

In functional magnetic resonance imaging, the brain's response to experimental cognitive tasks is usually assumed to be independent of endogenous oscillations. To test this assumption, we measured fractal scaling of fMRI time-series before and after a working memory task. Prolonged and task difficulty-related changes in post-task 'resting' data suggest that brain dynamics recover slowly from cognitive effort, contrary to the reflexive model that background oscillations are independent of task performance.

Highlights

The dominant experimental paradigm in functional magnetic resonance imaging assumes that the brain’s response to an experimentally controlled task is independent of its endogenous or background oscillatory activity
We addressed the possible interaction between cognitive task performance and endogenous functional magnetic resonance imaging (fMRI) oscillations in an experiment designed to answer two questions: 1) Does performance of a cognitively effortful task significantly change fractal scaling properties of fMRI time series compared to their values before task performance? 2) If so, can we relate the extent of taskrelated perturbation to the difficulty of the task and can we estimate how long it takes after task completion for the brain’s endogenous activity to recover or return to pre-task values?
There was no significant difference for mean reaction time (RT) between one-back and twoback working memory tasks (t(13) = 1.67, p = 0.06), there was a significant deterioration of accuracy (t(13) = 2.71, p = 9.3661023) at higher load

Summary

Introduction

The dominant experimental paradigm in functional magnetic resonance imaging (fMRI) assumes that the brain’s response to an experimentally controlled task is independent of its endogenous or background oscillatory activity. Most fMRI experiments are predicated on reflexive rather than adaptive models of brain function [1] This core assumption is reflected in many ways, including the design of experimental sessions and the use of linear models for time series analysis. It has been shown that low frequency fMRI oscillations have fractal scaling properties, in common with the electrocardiogram (ECG) and other physiological time series, that can be measured using the Hurst exponent and are modulated by normal ageing, Alzheimer’s disease and anticholinergic drug treatment [2,3]. In functional magnetic resonance imaging, the brain’s response to experimental manipulation is almost always assumed to be independent of endogenous oscillations. We addressed the possible interaction between cognitive task performance and endogenous fMRI oscillations in an experiment designed to answer two questions: 1) Does performance of a cognitively effortful task significantly change fractal scaling properties of fMRI time series compared to their values before task performance? 2) If so, can we relate the extent of task-related perturbation to the difficulty of the task?

Methods

Results

Conclusion