This paper reflects the research and thoughts of a student at the time the paper was written for a course at Bryn Mawr College. Like other materials on Serendip, it is not intended to be "authoritative" but rather to help others further develop their own explorations. Web links were active as of the time the paper was posted but are not updated.

Contribute Thoughts | Search Serendip for Other Papers | Serendip Home Page

Biology 202, Spring 2005 Second Web Papers On Serendip

Functional Magnetic Resonance Imaging (fMRI): Much Ado About What?

Camilla Culler

A cursory review of research in cognitive neuroscience reveals how widespread the use of neuroimaging technologies has become during the last ten years. Of these relatively new neuroimaging methods, functional magnetic resonance imaging (fMRI) has occupied the dominant role (1). As opposed to PET, which requires the use of radioactive markers that limit the frequency of exposures for a single participant, fMRI is a non-invasive method that purports to measure neural activity while a person engages in cognitive tasks (2). Unlike MRI which provides a static picture of the structure of the brain, fMRI provides both structural and functional images of the brain (3). The fMRI technique that has been used most frequently in cognitive neuroscience research is the BOLD (Blood Oxygen Level Dependent) method. The assumption that underlies BOLD is that neural activation is correlated with changes in blood flow and blood oxygenation and that the magnetic properties of oxygenated and deoxygenated blood are not the same (4). Banich (4) has referred to a "veritable explosion of research" that has appeared using the BOLD method of fMRI.

It would be hard to overstate the enthusiasm with which the neuroscientific community has embraced the use of fMRI methods. Gazzaniga and Heatherton (5) believe that fMRI has contributed to a "biological revolution" in psychology permitting scientists to begin to answer "some of the most central questions of human experience." Gazzaniga, Ivry, and Mangun (1) report that between 1998, when the first edition of their text appeared, and 2002, when the second edition was published, there have been an "explosion" of brain imaging studies with "dozens of dozens" of academic and clinical centers doing research with MRI scanners and close to 1500 brain imaging studies published for each of the four years. Posner and Raichle (3) refer to fMRI as the "most promising" of the new brain imaging technologies and Bear, Connors, and Paradiso (6) report that the advent of fMRI techniques has provided an "extraordinary opportunity" for scientists to witness the inner workings of the thinking, feeling, and responding brain.

But all is apparently not rosy in the neuroimaging world. There have been a growing number of cautionary voices that are beginning to appear in print that address the limitations of fMRI. Specific, technical concerns are about what conclusions can be made about the relation between fMRI and neuronal activity, and how neuronal activity, blood flow in the brain, and fMRI signals are connected (7). For example Heeger and Ress (7) point out the fMRI is an indirect measure of brain activation (brain cells firing) and address some of the weaknesses of the "linear transformation model". This model uses a mathematical formula to interpret fMRI results, and states that the strength of the fMRI signal is proportional to local neuronal activity that has been averaged over a space of several millimeters and over a time period of several seconds. Although Heeger and Ress conclude that the linear transformation model is a "reasonable and useful approximation" for what is actually taking place in the brain, they qualify this conclusion by stating that it applies to only some recording sites, in only some brain areas, and only when using select experimental protocols.

Heeger and Ress also cite several factors that may influence the relationship between the variables of the neural activity, the fMRI signal and the blood flow in the brain. These include the fMRI acquisition technique (BOLD results sometimes differ from other non-BOLD methods), the behavioral and stimulus protocols that are used (one working memory task may produce different fMRI results than another working memory task), the data analysis method that is used, and lastly how the neuronal activity itself is quantified and measured.

A recent article (8) refers to the "growing controversy over fMRI scans" and quotes several prominent neuroscientists who cite a range of concerns about the use of fMRI, the limitations of the method, and the reliability and validity of the conclusions that have been made on the basis of fMRI data. The initial excitement over fMRI and the great expectations may have been in part due to the fact that it does provide both spatial and temporal improvements in image resolution when compared with the older and more expensive PET scan (4). More than one researcher has referred to fMRI data as "gross" claiming that the localization of cognitive functioning is not consistent with the notion that brain activity for even simple cognitive activities is distributed in neural networks (8). This search for the localization of function has led some critics of fMRI to dismiss it as a 21st century variety of 19th century phrenology (8). There is also speculation that some of the false confidence in fMRI results may be due to the fact that the vivid, colorful graphics that fMRI produces, suggest a level of precision in measurement that is misleading (8).

Others have voiced concerns over fMRI's imaging power, the ability to make generalizations about individual brains when using fMRI data that are based on group averages, questionable forensic applications (using fMRI results as evidence in a court of law), as well as neuromarketing applications (using fMRI results to tell how consumers respond to certain products) that raise a host of neuroethical concerns (9). For example, in one study where six different people were given the same spatial memory task to perform, the fMRI scans for each of the six yielded extremely varied patterns of activation (8). In another study that looked at the findings of 38 different fMRI studies that purported to locate the region in the brain responsible for "executive functioning" the areas that were identified varied considerably from study to study (8). Therefore fMRI results cannot be generalized to entire populations, as each individual's result is different.

Recent efforts to create new lie detectors based on fMRI technology are being made while at the same time they have been subject to considerable criticism. Preliminary fMRI results show that when subjects lie, their anterior cingulate cortex and superior frontal gyrus are activated. Yet other studies show that activation of the anterior cingulate occurs when people are making decisions about a variety of things, not just whether to tell the truth or not (10). Langleben, who conducted these studies on fMRI and lie detection, concedes that the development of an fMRI lie detector that works effectively outside the controlled laboratory environment will take many years before it is realized (9).

These apparent contradictions mean that the use of fMRI is controversial. On the one hand, fMRI is touted as "revolutionary" and literally thousands of neuroscience studies during the past few years have made use of this method, while on the other hand, considerable criticism of fMRI has begun to appear. These criticisms, at the very least, call into question some of the conclusions that have been reached in studies making use of this technology, and suggest that the future of fMRI research and application is not so certain. What does fMRI reveal about brain-mind relations? What are its limitations? To what extent does the application of fMRI to forensics, the marketplace, etc., raise legitimate bioethical concerns? Although these questions, and related ones have yet to be completely answered, they raise serious concerns and present new challenges to neuroscientists. At least for now, "where fMRI is concerned, 'a penny for your thoughts' is currently more like 'a million pennies for a group-averaged hemodynamic response to highly constrained stimuli under entirely artificial conditions" (9).

References:
1) Gazzaniga, M., Ivry, R. & Mangun, G. Cognitive Neuroscience: The Biology of the Mind. New York: W.W. Norton, 2002.
2) Stirling, J. Cortical Functions. London: Rutledge, 2000.
3) Posner, M. & Raichel, M. Images of Mind. New York: W.H. Freeman, 1997.
4) Banich, M. Cognitive Neuroscience and Neuropsychology. New York: Houghton Mifflin, 2004.
5) Gazzaniga, M., & Heatherton, T. Psychological Science: Mind, Brain, and Behavior. New York: W.W. Norton, 2003.
6) Bear, M., Connors, B., & Paradiso, M. Neuroscience: Exploring The Brain. Baltimore: Lippincott Williams & Wilkins, 2001.
7) What Does fMRI Tell Us About Neuronal Activity?, an article by Heeger, D. & Ress, D. Nature Reviews/Neuroscience 2002, 3:142-151.
8) Fact Or Phrenology?, an article by Dobbs, D. Scientific American Mind, 2005, 16: 24-31
9)fMRI Beyond The Clinic: Will It Ever Be Ready for Prime Time?, an article by Robinson, R. PLOS Biology, 2004.
10) The New Lie Detectors , an article by Tancredi, L. Scientific American Mind, 2005, 16: 46-47e.


| Course Home | Serendip Home |

Send us your comments at Serendip

© by Serendip 1994- - Last Modified: Thursday, 19-Jul-2012 20:33:50 EDT