Villanova PSY 4200 - What Can Neuroimaging

Unformatted text preview:

What Can Neuroimaging Tell UsAbout the Mind?Insights From Prefrontal CortexRussell A. Poldrack1and Anthony D. Wagner21Department of Psychology and Brain Research Institute, University of California, Los Angeles, and2Department of Psychology and Neurosciences Program, Stanford UniversityABSTRACT—Psychologists interested in the workings of the mindmay wonder whether brain-imaging data can provide insightregarding cognitive mechanisms. Here we consider one meansthrough which imaging can inform cognitive theory: reverseinference, wherein activations in well-characterized neuralstructures serve as markers for the engagement of particularcognitive processes. To illustrate this approach, we reviewbrain-imaging evidence regarding the organization of cognitiveand linguistic processes in the prefrontal cortex, which indicatesthat phonological (speech-sound-based) and semantic (meaning-based) processing are consistently associated with topographi-cally distinct patterns of activity in the left inferior prefrontalcortex. We then illustrate how this finding of regional differen-tiation has provided useful guidance for understanding thecognitive processes supporting memory encoding and retrieval.We conclude with caveats that highlight some of the limitationsof the reverse-inference approach.KEYWORDS—cognitive control; semantic; phonological; neuroim-agingImaging of brain function has become an immensely popular tech-nique over the past decade, yet many readers of this journal mayharbor suspicions that it is nothing but a high-tech (and very expen-sive) revision of phrenology. A cursory glance at the neuroimagingliterature readily reveals instances of ‘‘blob-ology,’’ wherein re-searchers, often in a post hoc manner, attempt to explain why a par-ticular region was active during performance of a particular task. Yet,a broader consideration of the literature can reveal consistent patternsof activation that transcend the limitations of such ad hoc conclusions,appearing to reveal something more fundamental about the mappingbetween mind and brain. Such discoveries, which typically emergefrom an extensive body of investigation, may nevertheless still leaveopen the question: What is the use of knowing which brain regions areactive in association with a particular cognitive process?In this article, we aim to illustrate how knowledge of functionallocalization (i.e., the location of brain activations) can inform cogni-tive theories through the approach of reverse inference, wherein acti-vation in a particular brain region (or regions) is taken as a marker ofengagement of a particular cognitive process. We begin by describinga set of findings that suggest that meaning-based (semantic) andspeech-sound-based (phonological) processes are consistently asso-ciated with topographically distinct patterns of activity in the leftinferior prefrontal cortex (LIPC; see Fig. 1). We then illustrate howthis discovery of localized processes can serve as a useful (thoughlimited) guidepost, permitting neuroimaging data to inform cognitivemodels. We conclude by emphasizing some caveats regarding thereverse-inference approach.EVIDENCE FOR FUNCTIONAL SEGREGATION IN LIPCFunctional neuroimaging measures local brain activity indirectlythrough the imaging of such parameters as blood flow or blood oxy-genation. The effects of neurotransmitters released into the synapsesbetween neurons result in increased blood flow and oxygenation in theregion of those neurons, so that by comparing images acquired duringthe performance of tasks that differ in whether they engage a partic-ular cognitive process, one can determine which regions exhibit ac-tivity that is putatively related to that cognitive process.Some of the earliest neuroimaging studies of language revealedactivity in LIPC on task comparisons meant to isolate semanticprocessing, phonological processing, or both, and an extensive body ofsubsequent research supported the conclusion that this area of thebrain is activated during language production and comprehension.Tasks used to examine semantic processing have included generatingsemantic associates of cue words and classifying the meanings ofwords, such as deciding whether they are abstract (e.g., love)orconcrete (e.g., table). Tasks used to examine phonological processinghave included deciding whether words or pseudowords rhyme,counting syllables of words or pseudowords, and maintaining verbalmaterials in working memory (the immediately accessible form ofmemory in which information is held in mind and manipulated).Fiez (1997) suggested that tasks requiring semantic or phonologicalprocessing engage anatomically distinct regions in LIPC. ActivationAddress correspondence to Russell A. Poldrack, UCLA Departmentof Psychology, Franz Hall, Los Angeles, CA 90095-1563; e-mail:[email protected] DIRECTIONS IN PSYCHOLOGICAL SCIENCEVolume 13—Number 5 177Copyright r 2004 American Psychological Societyduring performance of semantic tasks was often observed in the moreanterior (forward) and ventral (downward) section of LIPC, whereasactivation during phonological tasks was often observed in the moreposterior (rearward) and dorsal (upward) section of LIPC (Fig. 1).Subsequently, an extensive meta-analysis (i.e., an analysis combiningthe results from a number of studies) provided further support for thisproposed functional segregation (Poldrack et al., 1999): Activation inanterior LIPC (aLIPC) was observed primarily during semantic tasks,whereas activation in posterior LIPC (pLIPC) was observed duringtasks requiring phonological judgments, as well as during somesemantic tasks. These findings led to the semantic-phonological hy-pothesis (SPH), according to which semantic and phonological pro-cesses differentially depend on the anterior and posterior subregionsof LIPC.Although early studies and meta-analyses motivated the SPH, di-rect evidence awaited within-subjects comparison of brain activationduring semantic and phonological processing. A number of studiespublished since 1999 demonstrated differences between aLIPC andpLIPC activity when semantic and phonological tasks were compareddirectly (e.g., Devlin, Matthews, & Rushworth, 2003; Poldrack et al.,1999). For example, Otten and Rugg (2001) contrasted activationpatterns when subjects made animacy judgments and syllable-counting judgments, observing greater aLIPC activation during theanimacy task than during the syllable-counting task, but greaterpLIPC activation during the


View Full Document

Villanova PSY 4200 - What Can Neuroimaging

Documents in this Course
Load more
Download What Can Neuroimaging
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view What Can Neuroimaging and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view What Can Neuroimaging 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?