Study design in fMRI: Basic principles

Abstract

There is a wide range of functional magnetic resonance imaging (fMRI) study designs available for the neuroscientist who wants to investigate cognition. In this manuscript we review some aspects of fMRI study design, including cognitive comparison strategies (factorial, parametric designs), and stimulus presentation possibilities (block, event-related, rapid event-related, mixed, and self-driven experiment designs) along with technical aspects, such as limitations of signal to noise ratio, spatial, and temporal resolution. We also discuss methods to deal with cases where scanning parameters become the limiting factor (parallel acquisitions, variable jittered designs, scanner acoustic noise strategies).

Introduction

Functional magnetic resonance imaging (fMRI) is a widely used technique to probe brain function, although the mechanisms underlying the information produced are not fully understood (Logothetis, Pauls, Augath, Trinath, & Oeltermann, 2001). An fMRI experiment depends upon techniques and methodologies derived from different fields of expertise, making it intrinsically multidisciplinary. From image acquisition to final data analysis, fMRI represents a challenge to the neuroscientist wishing to make the best use of the technique. It is therefore of the utmost importance to achieve some level of common understanding of the concepts involved in an fMRI project, to allow for efficient information exchange.

This manuscript is aimed at those not familiar with designing fMRI experiments, providing a framework for understanding the techniques available in the field and bringing together concepts otherwise only found scattered through the literature. Our focus will be on experiment design techniques, and we encourage the reader to refer to other excellent reviews for a broader view of fMRI in general (Matthews and Jezzard, 2004, Ramsey et al., 2002). There is a wide range of fMRI study designs available for the neuroscientist who wants to investigate cognition. In fact, the search for new methods is endless, and neuroscientists are often found in a state of helpless disappointment at the lack of ‘simple push button solutions’ in fMRI. In this article, we hope to clarify some aspects of the technique, describing the major factors influencing the measured signal, introducing a practical view of cognitive comparison strategies, describing conventional acquisition schemes (i.e., block designs, event-related) and new ‘self driven’ approaches, and introducing common issues in fMRI studies. We briefly mention special cases, such as problems related to the effects of acoustic noise from the scanner, and other technicalities, such as limitations in the signal to noise ratio (SNR) and the spatial and temporal resolution of the method, providing an introduction to the major concepts inherent in the field. Next, we highlight the main issues that emerge when trying to integrate ‘smart’ cognitive comparison strategies (factorial, parametric designs) with ‘limited’ scanning parameters like brain coverage and temporal resolution (variable jittered designs, parallel acquisitions). Finally, as selecting the correct combination of strategies in each aspect of the technique is crucial to interpretation of the results, we encourage the use of study designs with the minimum degree of complexity possible.