Experimental design and interpretation of functional neuroimaging studies of cognitive processes

David Caplan
2009 Human Brain Mapping  
This article discusses how the relation between experimental and baseline conditions in functional neuroimaging studies affects the conclusions that can be drawn from a study about the neural correlates of components of the cognitive system and about the nature and organization of those components. I argue that certain designs in common use-in particular the contrast of qualitatively different representations that are processed at parallel stages of a functional architecture-can never identify
more » ... he neural basis of a cognitive operation and have limited use in providing information about the nature of cognitive systems. Other types of designs-such as ones that contrast representations that are computed in immediately sequential processing steps and ones that contrast qualitatively similar representations that are parametrically related within a single processing stage-are more easily interpreted. basis of cognitive functions have been made at all levels of detail, from specific operations through components. The study of the neural basis of cognitive operations involves contrasting tasks that are assumed to engage operations or components that are related to one another in ways that are specified in a functional architecture. This article considers the conclusions that can be drawn from particular contrasts. It reviews seven functional architectures and argues that contrasts of operations that are related in particular ways can, in principle, isolate cognitive operations or components, and that contrasts of operations that are related in other ways cannot, in principle, do so. 1 Though the examples of studies are drawn from functional magnetic resonance imaging, the conclusions of this article apply to all forms of neuroimaging that use the designs discussed here, including event related potentials, magnetoencephalography, optical imaging, and others. 2 PRELIMINARY ISSUES A few preliminary issues pertaining to how cognitive processes are related to neural observations are necessary before turning to the discussion of experimental designs. Cognitive operations require "processing resources" [Shallice, 1988 [Shallice, , 2003 . Different operations and sets of operations make greater or lesser demands on processing resources, leading to changes in responses such as accuracy, reaction time, BOLD signal level, amplitude and latency of electrophysiological signals, etc. Two fundamental assumptions about neural activity are that (1) measures of neural activity such as BOLD signal, electrophysiological signals measured at the scalp, etc., are related to cellular and subcellular neural events that are the basis for cognitive functions; and (2) as a cognitive task requires more operations, or more "effort" or "processing resources" to apply an operation, the measured neural activity increases. 3 1 An analogy to syllogisms may in some ways capture the difference between the issues discussed here and those that arise in connection with individual studies. Syllogisms can be valid (i.e., the conclusion follows from the premises) and/or sound (i.e., the premises are true). Some syllogistic figures cannot yield a valid conclusion no matter what the truth of their premises; e.g., the premises Some carpenters are plumbers and some plumbers are electricians cannot yield a valid conclusion relating carpenters and electricians even if the premises are true. Syllogisms with false premises can yield valid conclusions, but the conclusions are likely to be false; e.g., the premises All carpenters are plumbers and all plumbers are electricians yields the valid conclusion that all carpenters are electricians, but this conclusion is likely to be false because both of the premises are false. Like syllogisms, studies whose designs could, in principle, yield interpretable results can be considered valid. However, such designs can still fail to provide data about the operations they purport to study because of flaws in the choice of materials and tasks (on analogy to syllogisms, they are not sound). 2 The issues discussed in this article are relevant to studies that use functional neuroimaging to study the neural basis of cognition and the organization of cognitive processes. They are not directly relevant to studies of the neural correlates of cognitive functions that are not directed at the question of how the brain is organized to perform a cognitive operation. For instance, a study could identify different neurovascular responses to a cognitive task in teenagers with and without ADD. This might be useful in many ways (e.g., in making a diagnosis in unclear cases) even if the task did not provide information about how the brain supported a cognitive function. However, if a researcher wants to know whether there is a difference in the way the brains of teenagers with and without ADD perform a particular cognitive function, such as dividing attention, it becomes necessary to use experimental designs that could identify brain regions that support those functions. In that case, the issues discussed here become important to consider. 3 These assumptions have been questioned; this note is designed to answer some of these concerns. Page [2006] has argued that Logothetis et al. [2001] showed that BOLD signal is not a measure of axonal spike activity, and thus not a measure of neuronal activity relevant to information processing since spike activity is the efferent signal of a brain area. This, however, misconstrues Logothetis et al. [2001]. Logothetis et al. [2001] found that BOLD signal correlated more highly with measures of dendritic activity (local field potentials) than with axonal spike frequencies, not that the correlation of BOLD signal and axonal spike frequency was not significant. It also bears noting that dendritic activity is an electrochemical state of a brain region that is likely to be informationally-relevant. Page [2006] has also argued that increased in neurovascular responses might reflect increased inhibitory functioning of an area of the brain. This is unlikely: the percentage of GABAergic inhibitory neurons cortical neurons is estimated at between 15 and 25% [Kisvarday et al., 1990 ]. It is not now known whether there are areas of cortex with higher percentages of inhibitory neurons. If there are, BOLD signal increases in these areas can be interpreted accordingly. Finally, there are several reasons to believe that increased BOLD signal reflects increased cognitive demand [see Shallice, 2003, for discussion]: linear models are good fits to BOLD signal [Boynton et al., 1996] ; stimuli that affect behavioral measures in ways that indicate that they are more demanding lead to increases in BOLD signal; as tasks become easier, more practiced, and more automatic, neurovascular activity decreases [Raichle et al., 1994] . It bears mention that the relation between behavioral DVs and IVs is not always monotonic [the Yerkes-Dodson Law; Kahnemann, 1973] or linear [McClelland, 1979] but it is possible to use stimuli whose behavioral effects are known to be linear in the range of the IV that is being manipulated. Caplan
doi:10.1002/hbm.20489 pmid:17979122 pmcid:PMC2612094 fatcat:vmmyl3bqpzd4zbdm3pi33fh24q