心理学报 2010, Vol. 42, No.1, 111−119
杂酚油Acta Psychologica Sinica DOI: 10.3724/SP.J.1041.2010.00111 111
The Value of Brain Imaging in Psychological Rearch
Richard Gonzalez and Marc G. Berman
(Department of Psychology, University of Michigan, USA)
Abstract: We take the view that commonly ud brain imaging techniques add new and informative data to psychological rearch. As with any new measure, we need to decide how to u it in an appropriate way. How does the measure help answer theoretical questions in ways that existing measures cannot? Is the measure best ud as a dependent variable or as a predictor variable? How does it relate to other psychological variables of interest? This new imaging technology provides exciting glimps into the workings of the brain and its relation to psychology. Rearchers need to figure out how the information provided can be ud to advance the understanding of psychological phenomena. Key words: fMRI; statistical modeling; psychological theory
Introduction
Functional brain imaging offers the ability to examine a person’s brain while that individual engages in a psychological activity of interest. The promi of brain imaging techniques is that they permit tracking the brain in real time. They provide the rearcher the opportunity to look “under the hood.” In this paper, we consider how such information is uful in addressing psychological rearch questions, including how anatomical localization can relate to psychological process and how dissociation can provide a test of underlying psychological process. We write from the perspective of a traditional psychologist interested in learning more about what the techniques offer. We provide advice about how one can add the techniques to a rearch program to maximize the chance of uful discovery and facilitate incremental knowledge. We limit our attention to the broad category of brain imaging techniques with special attention to functional magnetic resonance imaging (fMRI).
Received date: 2009-06-11
Correspondence concerning this article should be addresd to Richard Gonzalez, E-mail: gonzo@umich.edu
越狱大结局
Background: It is all about the measure
Brain imaging techniques measure different variables. Electroencephalograph (EEG) measures elect
rical activity in the cortex, magnetoencephalography (MEG) measures magnetic fields that are biproducts of electrical activity in the cortex, fMRI measures oxygen levels in blood, and voxel-bad morphology of magnetic resonance (MR) images measures cortical thickness. The different techniques can be construed as different measured variables. Their u in scientific rearch becomes relevant as long as the variable is related to the underlying psychological process under study. As with any empirical science the paradigm includes measured variables, which can be formulated as either dependent or predictor variables in theoretical and statistical models; the value of tho measures is driven mostly by the scientific advances they facilitate. The basic thesis of this paper is the following: brain imaging prents a new measured variable and the value of such a measurement is driven not by any provocative construal we provide about “brain activation” and colorful images of the brain that “light up” but by the scientific advances, the theoretical developments, and the empirical testing afforded by such measurements. This is true of any
维吾尔语翻译>rig112 心理学报 42卷
measured variable ud in empirical rearch. Just becau a technology produces colorful images of the brain does not make it more scientific than any other measure ud in psychology. Our tone may em cautious and reactionary but we are responding more to what we believe are fal claims
and fal hopes that rearchers place on brain imaging, especially tho new to imaging techniques.
Some have criticized recent rearch using brain imaging as simply being modern phrenology becau it appears, to the untrained eye at least, that the goal of the program is to map and localize brain function (Uttal, 2001). Fear is in the amygdala, social psychological processing is in the medial prefrontal cortex, memories are formed in the hippocampus. Every process has a place, only now rather than probing the peaks and valleys on the scalp as did the early phrenologists, the modern scientist has access to an expensive electric magnet that produces data that can be reconstructed into images and statistical maps. We agree that some modern brain imaging contains some aspects of localization on par with the early phrenologists, but we also believe that imaging techniques hold much promi when viewed as another variable in the psychologist’s toolbox. Many rearchers have utilized the technology well beyond simple “phrenology.” For example, Kosslyn, Thompson, and Alpert (1997) ud Positron Emission Tomography (PET) to show the shared systems of imagery and visual perception, thus demonstrating the similarity of tho psychological process. Raichle, MacLeod, Snyder, Powers, Gusnard, and Shulman (2001) ud brain imaging to provide novel insights into the baline state of the human brain, helping us learn more about the psychological pr
出埃及记马克西姆钢琴曲mp3下载ocess involved in “off-task behavior.” In both cas, neuroimaging was ud beyond localization and provided insights into psychological process that would be difficult to uncover with behavioral data alone.
Additionally, localization is typically not the primary goal of brain imaging, nor should it be. Our field has deep psychological questions that go beyond where a specific psychological activity is located in the brain. Further, localization may play an important role in developing theories about a network of brain respons and how brain structures work as a system to take multiple stimuli and multiple internal process to produce multiple behaviors, cognitions and emotions. Thus, localization is one u of brain imaging techniques, but it is not the only u, nor necessarily the best u. We acknowledge, though, that localization knowledge may be uful as we develop and test psychological models. Localization can potentially help to classify a t of cognitions and actions that may be difficult to classify with behavioral experimentation alone. However, we must not constrain our thinking and view brain imaging techniques as limited to only localization. The techniques provide measured variables and their value depends on how we u them in our rearch, whether the measured variables allow us to test burning rearch questions, and whether the variables prompt us ask new rearch questions and propo new experimental tests.
This view parallels the u of reaction time measures in psychology. That one psychological process takes longer than another is not necessarily uful information in isolation, but ud in a clever way reaction time measures can provide new tests of psychological process that were not readily available with other measures. For example, Sternberg (1966) ud reaction time to demonstrate that recognition time of an item in short term memory depends in part on the number of items in storage. Luce (1986) reviewed the rich t of reaction time studies on detection, identification and matching paradigms, along with mathematical models, showing the deep lessons that can be learned about cognition by knowing how to work with a dependent variable. The examples illustrate that reaction time can provide information beyond which condition yields faster respon. The localization that emerges in imaging studies should analogously be viewed as information that can help
1期The Value of Brain Imaging in Psychological Rearch 113
us make inferences about psychological process. In short, we do not view localization information as the best method for making psychological inferences nor should it be the sole end goal of an imaging study. We will discuss other us of fMRI that extend beyond localization, which will be beneficial to psychological theorizing.
reboot是什么意思
Different Technologies
There are different technologies that can be ud in brain imaging studies. It is helpful to categorize the techniques along a t of dimensions that relate to psychological questions. One dimension is brain function versus brain structure. Are we interested in imaging the brain during psychological activity (functional) or are we interested in morphological properties such as the thickness of the cortex or the fidelity of white matter connections that form the major routes of neural activity (structural). Some rearch questions, such as tho involving age-related changes in cognition, may invoke both structural and functional rearch questions so more than one imaging technique may be relevant. Are there any structural differences in the brain between older and younger adults? Are there functional brain differences that relate to behavioral differences between older and younger adults? Are there age differences in brain networks?
Another dimension is the relative importance of spatial versus temporal resolution. Do we want to optimize our ability to localize specific brain activation to a particular spatial locale or do we want to optimize our ability to measure when brain activation occurs at the expen of not knowing the preci location of that activity? This dichotomy is not due to some theoretical limit related to some uncertainty principle in physics, rather existing technology does not permit both high spatial and tem
poral resolution concurrently in the same technique, though this is an active area of rearch and some promising new developments have been made. Some rearchers have been successful at using two methodologies concurrently (such as EEG and fMRI) in order to realize both spatial and temporal resolution (e.g., Goldman, Stern, Engel, & Cohen, 2000).
Rearch questions
Many of us are not aware of the extent to which the particular paradigm or methodology we u can constrain the types of rearch questions we ask. This usually emerges early in our psychological training as young graduate students when we are socialized into a particular rearch lab, a particular way of doing science, a particular rearch literature, a t of friends we cite and foes we ek to dismantle with our creative studies. If we commonly u reaction time data, then we tend to ask questions that can be tested by a reaction time paradigm; if we commonly u lf-report data, then we ask rearch questions that lend themlves to testing with lf-report data. The reader may be familiar with the “culture shock” associated with entering a new domain of rearch where “things are done differently,” perhaps a different measure is taken, or a different paradigm reigns supreme, or a different type of stimulus is ud. Sometimes our reaction to the new is skepticism becau it ems messy and foreign to us; the rituals about how to process the data or how to test subjects e
m ad hoc. Some rearchers retreat back to the safe known world of what was learned in graduate school, but others venture out to tackle the unknown. In doing so we may come to realize that what was so natural and straightforward to us is also in a n ad hoc and full of ritual.
Moving a traditional psychological laboratory to incorporate brain imagining techniques carries with it an analogous t of issues on at least two levels. First, the psychologist must learn relevant neuroscience. It is not sufficient to work at the level of the psychological theory and invoke psychological constructs. Rather, one must learn a little physiology, a little neuroanatomy, and a little biopsychology. This is not easy becau it involves
114 心理学报 42卷
learning new fields, not merely new areas within a field. However, doing so will pay great dividends, and lead to much more coherent imaging rearch with relevance to psychological theory. Second, the psychologist must learn the relevant details of the brain imaging technique. If one wants to study positive emotions in the fMRI scanner, how does being confined in a tube emitting a loud jackhammer-type noi influence the type of emotion that a subject could possibly experience? An imaging design requires the paradigm to keep track of various ont times, with psychological proce
ssing happening within, say, 15-30 cond trials, but if our usual social psychological paradigm involves participants reading a long paragraph and then answering five open-ended questions, there will be a mismatch –– the psychological paradigm will have to be re-worked
西安哪里学托福最好to match the constraints of the interpretable information the fMRI scanner can provide. Of cour, if a new paradigm is developed for u in the scanner, the careful rearcher should also replicate a few standard results in the field to verify that the new paradigm still yields the same underlying behavioral conclusions. In other words, the rearcher needs to demonstrate that the new “neuroimaging ready” paradigm produces results consistent with tho in the literature. Therefore, understanding the rearch context, the methodology, the limitations of the method, along with the constraints that the methodology places on types of rearch questions that can be asked, are important considerations when adding brain imaging to one’s psychological rearch program. One should not merely hire a technical person to handle the details of developing the fMRI paradigm. The psychologist needs to be actively involved in guiding the rearch paradigm so it maintains the fidelity of the psychological question.
Localization, Association and Dissociation
There are veral types of hypothes that can be addresd with brain imaging rearch. One is localization, which amounts to mapping psychological function to anatomical structures. While this may be an important type of question, in our view it is one of the least interesting in terms of providing new psychological understanding about process (e also Cacioppo et al, 2003). It isn’t clear that knowing that social anxiety is located in one place in the brain or another adds much to our body of knowledge, other than perhaps a uful boundary condition or another way to classify a psychological process. This qualification about location, however, mostly applies to the current type of location information offered by existing brain imaging techniques (e.g., blood flow). It is possible that advances in brain imaging technology and methodology could provide more uful localization information, such as providing information about neurotransmitters and neural circuitry –– esntially localization involving detailed information about the underlying neural mechanisms. Current types of atlas-style localization (e.g., amygdala, anterior cingulate cortex, medial prefrontal cortex) are too gross to pinpoint underlying neural mechanisms given that there are so many substructures (or layers) within structures, different types of transmitter receptors, etc.
fcaThere are other types of rearch questions that can be asked with brain imaging techniques that can lead to deep and unique information about psychological process. One such rearch question i
nvolves association. Somewhat analogous to converget validity, the types of rearch questions relate two or more process to each other. An example is the association between brain activity as measured by BOLD and subquent choice in a decision making task, or the association of multiple brain structures working in tandem as a network. Another example, is the one we cited earlier about the similarity between imagery and perception as found by Kosslyn et al. (1997).
A third type of rearch question involves dissociation. Dissociation is analogous to discriminant validity. If we are investigating two psychological process we can examine whether they are differentially driven by anatomical
1期The Value of Brain Imaging in Psychological Rearch 115
structures. We may be able to dissociate, say, the emotion of regret (which involves counterfactual thinking) from the emotion of disappointment (which involves an evaluation of the current state); e, for example, Chua, Gonzalez, Taylor, Welsh & Liberzon, 2009). Behaviorally, it is not easy to distinguish regret from disappointment becau they both can be strong negative emotions in respon to a negative decision outcome. Brain imaging data, for instance, can help dissociate the two process by finding brain regions that either respond differently in the two states or res
pond differentially to different experimental manipulations. For similar distinctions between different types of rearch questions that can be asked in a brain imaging context e Berman, Jonides, & Nee (2006) and Knutson & Greer (2008).
Neuroimaging Environment
One needs to lect phenomena that can be studied with the chon brain imaging technique, but be mindful of the neuroimaging environment, which is more confining than the typical laboratory tting. If one wants to study the gut level reaction that emerges when eing a disgusting scene, MRI may not provide the best temporal resolution for that psychological process. One may have to alter a tried and true familiar paradigm to fit the constraints of a brain imaging study. The usual social psychological or decision making paradigm of having participants read a paragraph scenario may not work in the fMRI scanner becau the brain activity involved in reading and comprehending a paragraph makes it difficult to define when a psychological event occurs if it is not known when particular brain activation should show up. Or, if one wants to study positive emotion, the context of the loud confining fMRI tube may swamp any positive mood induction created by the most well-intentioned behavioral paradigm. If one is studying xual attraction, then the rearcher must be mindful that the skull cap in the EEG study may make the rearch subject feel unattractive to the re
英文面试自我介绍
arch assistant of the opposite x, and may lead to an emotional state that contaminates the study. Scanning children creates a new host of challenges given that the child must stay still for a relatively long time. In many cas it is difficult to maintain a child’s interest throughout many trials of a relatively boring psychological study, let alone doing the task inside an MRI scanner.
Design
As stated previously, the type of design one us may need to be adapted to fit into a brain imaging context. One may be accustomed to using between-subjects manipulations in a behavioral tting, but a within-subjects design may be more efficient in a brain imaging context. This means revamping a traditional behavioral paradigm, which may open up new issues as one switches the details of the paradigm to a within-subjects design (e.g., Greenwald, 1976). In order to have sufficient statistical power in such psychological domains, rearchers simply test more subjects. Financial and time constraints may lower the practicality of this solution in an imaging context, and rearchers may have to develop new designs and manipulations that can be adapted to the brain imaging technique. This is especially important considering the amount of noi in neuroimaging data (e.g., Parrish, Gitelman, LaBar & Mesulam, 2000).至死不渝的近义词
Another important issue of design involves the particular method of prenting the experimental manipulation and the experimental stimuli. The rearcher must decide whether a block design is feasible or whether an event-related design is more appropriate for the type of rearch question being tested. A block design involves a quence of similar experimental trials organized in common “blocks” of trials (e.g., ten happy trials, followed by ten sad trials, or a t of high frequency word trials followed by a t of low frequency word trials), which optimizes signal, but does not allow the modeling of individual trials. An event-related design involves a design where experimental trials are interleaved (e.g., a happy trial, followed by two sad trials, followed by a happy trial, etc, where
