Transferring a paradigm with a history of use in EEG experiments to an fMRI experiment is considered. It is demonstrated that manipulating the task demands in the visual oddball task resulted in different patterns of BOLD activation and illustrated how task design is crucial in fMRI experiments.
As cognitive neuroscience methods develop, established experimental tasks are used with emerging brain imaging modalities. Here transferring a paradigm (the visual oddball task) with a long history of behavioral and electroencephalography (EEG) experiments to a functional magnetic resonance imaging (fMRI) experiment is considered. The aims of this paper are to briefly describe fMRI and when its use is appropriate in cognitive neuroscience; illustrate how task design can influence the results of an fMRI experiment, particularly when that task is borrowed from another imaging modality; explain the practical aspects of performing an fMRI experiment. It is demonstrated that manipulating the task demands in the visual oddball task results in different patterns of blood oxygen level dependent (BOLD) activation. The nature of the fMRI BOLD measure means that many brain regions are found to be active in a particular task. Determining the functions of these areas of activation is very much dependent on task design and analysis. The complex nature of many fMRI tasks means that the details of the task and its requirements need careful consideration when interpreting data. The data show that this is particularly important in those tasks relying on a motor response as well as cognitive elements and that covert and overt responses should be considered where possible. Furthermore, the data show that transferring an EEG paradigm to an fMRI experiment needs careful consideration and it cannot be assumed that the same paradigm will work equally well across imaging modalities. It is therefore recommended that the design of an fMRI study is pilot tested behaviorally to establish the effects of interest and then pilot tested in the fMRI environment to ensure appropriate design, implementation and analysis for the effects of interest.
As cognitive neuroscience methods develop, established experimental tasks are used with emerging brain imaging modalities. This is a logical progression since most neuropsychological concepts (e.g., distinct memory sub-components) have been investigated in the behavioral domain and appropriate experimental tasks for probing specific functions have been developed and tested. As new technology emerges evidence for the neural underpinnings of these behavioral observations is sought with the new brain imaging methods. While it may be tempting to simply draw on well-studied behavioral tasks for imaging studies, several important caveats have to be taken into account. One crucial, though frequently neglected, consideration is the use of the most appropriate imaging technique to further probe the behavioral evidence. In terms of cognitive neuroscience and psychology there are many brain imaging methods available to enhance our understanding of the neural activity underlying concepts of interest; for example electroencephalography (EEG), magnetoencephalography (MEG), transcranial magnetic stimulation (TMS), functional magnetic resonance imaging (fMRI) and positron emission tomography (PET). All of these methods have their advantages, disadvantages and appropriate applications. Here transferring a paradigm with a long history of behavioral and EEG experiments to an fMRI experiment is considered. EEG has been used for decades to investigate neural responses associated with perceptual and cognitive processes. As such, many paradigms have been developed for use with this method and have evolved over time. Functional MRI is a technique that emerged more recently in cognitive neuroscience and this has led to some paradigms developed in EEG research being used in fMRI. To build on the knowledge base from EEG experiments with the new techniques is a logical step but nonetheless some important points may be neglected in the transfer. The techniques are very different and tasks need to be designed accordingly. This requires knowledge of how the method works and, in particular, how potential modulations of the paradigm used will influence the measures taken. For further information on the design of fMRI experiments the interested reader is directed to the following link http://imaging.mrc-cbu.cam.ac.uk/imaging/DesignEfficiency. Task design will be considered in the context of transferring a paradigm developed for EEG research to the fMRI environment. The aims of this paper are: i) to briefly describe fMRI and when its use is appropriate in cognitive neuroscience; ii) to illustrate how task design can influence the results of an fMRI experiment, particularly when that task is borrowed from another imaging modality; and iii) to explain the practical aspects of performing an fMRI experiment.
Functional MRI is now a widely available technique and as such is a common method used in cognitive neuroscience. In order to make a decision as to whether the technique is appropriate for a particular experiment the advantages and disadvantages of fMRI must be considered in relation to other available techniques. A disadvantage of the method is that it is not a direct measure of neural activity, rather it is a correlate of neural activity in that the metabolic response (oxygen requirement) convolved with the haemodynamic response. Thus its temporal resolution is poor in comparison to electrophysiology, for example, where the measured electrical signal is closer to the underlying neural activity rather than a metabolic response. EEG has a temporal resolution in the order of milliseconds compared to a resolution in the order of seconds in fMRI. However, the main advantage of fMRI is that the spatial resolution of the technique is excellent. Furthermore, it is noninvasive and thus subjects do not have to ingest substances such as contrast agents or be exposed to radiation as would be the case in positron emission tomography (PET). Therefore, fMRI is a suitable technique for experiments investigating which brain regions are involved in perception, cognition, and behavior.
In this paper the visual oddball paradigm is taken as an example for the transfer of a well-established EEG-task to fMRI (see Figure 1 for details). It should be noted that the issues discussed could also influence results and data interpretation when other paradigms are used and should technically be considered in the design of all fMRI experiments. The oddball paradigm is frequently used in psychology and cognitive neuroscience to assess attention and target detection performance. The paradigm was developed in EEG research, specifically event related potentials (ERPs), for investigating the so called P300 component1. The P300 represents target detection and is elicited upon recognition of an infrequent target stimulus1. The P300 is used in studies across a number of cognitive and clinical domains2 e.g., patients with schizophrenia and their relatives3, heavy smokers4 and the aging population5. Given that the oddball paradigm (and the P300 elicited by the paradigm) is robust and is also modulated by different disease states, its transfer across different imaging modalities was inevitable.
The widespread activation seen in the brain during an oddball fMRI measurement is known to be the result of multiple cognitive functions, as shown by numerous fMRI studies probing other cognitive concepts. This widespread nature of the activation pattern makes it difficult to determine which brain regions are more (or less) active due to the specific task manipulations or group differences that the experimenter is interested in. Specifically, it is not certain whether observed differences in activation are related to target detection itself, to attention related processes, or whether they are related to other task demands such as ongoing working memory processes or processes related to the production of a motor response. The process of assigning function to the measured activity is easier in the EEG domain where the cognitive component of interest (target detection) is measured in clear cerebral response to the oddball task (P300). Nevertheless, neuroscientists tend to interpret their findings in favor of their own hypothesis and experiment, rather than putting in the effort to rule out alternative explanations. Most experiments, however, will not be able to solve these important questions inherently – scan time is costly – which is why we argue for thorough planning and pilot testing of paradigms.
Besides this difficulty in establishing a direct link between brain regions and cognitive components, the nature of the oddball paradigm also presents other possible methodological issues when being transferred to fMRI. For example, the detection of a target stimulus is usually indicated by pressing a response button. This allows the experimenter to record the accuracy and speed of responses but this response may also impact on the fMRI BOLD response to target stimuli. The motor action required for the button press impacts on stimulus-locked fMRI activation given that it happens just a few hundred milliseconds after the presentation of the target stimulus. This can also influence interpretation of that activation, for example brain regions involved in preparing for the motor response might wrongly be assumed to be involved in the detection of the target stimulus, and vice versa. This has led to methodological modifications whereby indirect measures of target detection, not relying on motor responses, are taken. For example, counting target stimuli has been proposed6 as a way to make sure subjects maintain attention on the task; the number of trials missed can indicate how inattentive a subject was. Reporting the number of stimuli counted at the end of the task also means that the experimenter can check whether the subject performed the task correctly. A third alternative is to use a fully passive task design where the subject is given no instructions on how to respond and the novelty of a target stimulus is assumed to inherently elicit a target detection-like response. Despite these versions of the task using the same type of stimuli and basic design, the activation pattern resulting from each variation of the task will be different because the cognitive and motor demands of the tasks are different7,8. For example, there will be working memory processes involved in counting target stimuli e.g., holding the current number of target stimuli in mind, that will not be required during passive viewing. Here these 3 versions of the oddball task, passive, count ,and respond are used to show how careful task design and implementation can account for these changes in task requirements and allow appropriate interpretation of the results.
We show that manipulating the task demands in the visual oddball task results in different patterns of BOLD activation in the count and respond conditions. The functional roles of some of the regions implicated in each condition would have been inappropriately assigned had data from the three versions of the task not been available for comparison. This ambiguity in data interpretation would not necessarily have been the case in the EEG P300 field where the task has its origin, highlighting the need for …
The authors have nothing to disclose.
Name of Material/ Equipment | Company | Catalog Number | Comments/Description |
Magnetom Tim Trio 3T MRI scanner | Siemens Medical Solutions, Erlangen, Germany | ||
Presentation version 14.8 | Neurobehavioural system, Albany, CA, USA | ||
Lumitouch device | Photon Control Inc, Burnaby, BC, Canada | This device is no longer produced by the manufacturer. Alternative MR compatible response devices are available | |
TFT display | Apple, Cupertino, CA, USA | 30inch cinema display | The screen was custom modified in-house to be MR compatible. However, a number of MR compatible screens are available on the market |
optseq | surfer.nmr.mgh.harvard.edu/optseq | program for determining optimal stimulus timing for rapid event related designs | |
FMRIB software library (FSL) | FMRIB, Oxford | http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/ | Other software tools are available for analysing fMRI data, for example SPM, AFNI and Brain Voyager |