Custom Search
|
|
|
|
BACKGROUND: fMRI and the n-back task
In 1999 and 2001, Dr. Thomas McAllister published two fMRI (functional magnetic resonance imaging) studies with patients who suffered mild traumatic brain injuries. For those who are unaware, fMRI is an imaging technique that indirectly measures blood flow and neural (i.e., brain cell) activity. Both of McAllister’s studies involved presenting MTBI patients and healthy controls a test of working memory while the fMRI technique was performed. There are various definitions of working memory, but it generally refers to the ability to briefly register information, briefly maintain it, and perform some type of mental manipulation with it. An example would be saying a series of numbers backwards. The person needs to register the numbers, briefly maintain them, and reverse (i.e., manipulate) them.
In McAllister’s studies, the subjects performed a working memory task known as the n-back which is commonly used in fMRI studies. The first part of the n-back, known as the 0-back, does not actually measure working memory but measures vigilance. The task requires the subject to press a button each time a certain letter appears on the screen, such as an X. The next part of the n-back is the working memory measure. It is known as the 1-back. The task requires the subject to press a button if the letter on the screen was the same as the letter that immediately came before it. So if the subject saw the letters R, G, N, T, R, R, L, M, N, N, the subject would press the button the 2nd time the letters “R” and “N” were presented. On the 2-back, the subject is required to press the button if the letter on the screen was the same as the letter that which came two places before it. So if the subject saw the letters R, G, N, T, R, R, R, L, M, N, N, N the subject would press the button the 3rd time the letters “R” and “N” were presented.
Technically, one could administer a 4-back, 5-back, 6-back, etc., but researchers typically do not exceed the 3-back task due to increased task difficulty. By subtracting the brain activity level on the 0-back (non-working memory task) from the 1-back, 2-back, and 3-back tasks, the researcher controls for activity involved in pressing the button and paying attention to stimuli on the screen. The remaining activity is presumed to reflect the activity from the working memory component of the task. The researcher can also subtract the activity of the 1-back from the 2-back (and of the 2-back from the 3-back) to determine how much more activation occurs by increasing the processing load.
THE STUDIES
In McAllister’s 1999 study, he used the 0-back, 1-back, and 2-back tasks. In his 2001 study, he added a 3-back task. In the 1999 study, McAllister showed that when compared to controls, MTBI subjects used significantly less brain activation when increasing from the 0-back to the 1-back but more brain activation when increasing from the 1-back to the 2-back. In McAllister’s own words, “The explanation for this difference in activation pattern between patients and control subjects is not immediately obvious” (p. 1304). However, because the MTBI subjects complained of more memory problems than controls, McAllister hypothesized that the fMRI activation differences might be related to these subjective complaints. In McAllister’s words, the results suggest “…that an alteration in the ability to activate or to allocate processing resources in response to a moderate working memory task may be associated with cognitive complaints after very mild TBI ” (p. 1307). In explaining why the MTBI patients showed less activation when increasing from the 0-back to the 1-back task, he suggested that this may reflect a difficulty that MTBI patients have in switching on their cognitive network.
McAllister did note that his interpretations should be tempered. It is also noteworthy that that the actual performance of MTBI subjects compared to controls did not differ on actual memory tests (verbal and visual-spatial) outside the scanner. These tasks were separate from the n-back. Although McAllister noted in his 1999 paper that the next paper would involve studying the same patients one year post-injury, this paper has not been published according to a search on PubMed.
In the 2001 study, McAllister again demonstrated a more significant increase in brain activation from the 2-back to the 1-back task and significantly less activation from the 2-back to the 3-back task, when MTBI patients were compared to controls. McAllister hypothesized that once the processing load reached a certain threshold (i.e., 3-back), that the MTBI patients had an impaired ability to match cognitive resources to the processing demands of a task, once a certain threshold had been crossed. However, he again acknowledged that the neural basis of the findings were unclear. Once again, there were no differences in actual performance between MTBI patients and controls. An image frequently reproduced to demonstrate these findings is displayed below.
THE PROBLEMS
The biggest misconception I have found about these studies is that the patients were studied more than three months to more than a year post-injury, with one person telling me that there were subjects in the study up to two years post-injury. One reason this myth may have emerged is because discussion of these papers frequently occurs in the context of discussions of patients who complain of symptoms more than three months post-injury. As a result, people seem to have automatically assumed that these studies apply to patients at that point in the post-injury time frame.
The facts are very different. First, one only needs to look at the title of the 1999 article to see that this was a study performed in patients one month post-injury. In fact, even that title is somewhat misleading as the study shows that the subjects were actually assessed within one month post-injury. Specifically, the average subject was studied 10.5 days post-injury, with the range being 6 to 35 days. The 2001 study also involved patients assessed within one month post-injury. Specifically, the average subject was studied 26.9 days post-injury, with no time range reported. Therefore, from the data that is available, the longest time post-injury that an MTBI subject was studied was 35 days, not two years.
The other problem is that this study has been used by some to suggest that normal neuropsychological test results during an outpatient clinical evaluation should be ignored in MTBI patients because these patients have differing patterns of brain activation that explain their subjective complaints. In other words, the argument would be something like the following: “Yes, the patient performed in the normal to superior range on all of the neuropsychological tests but because they used different amounts of brain activation, this explains why he/she may be exhausted afterwards and cannot function in everyday life.” Such arguments are overgeneralizations from the studies. Generally, MTBI patients are seen for outpatient neuropsychological evaluations more than three months post-injury. However, once this point has been reached, the patient is no longer in the acute or subacute injury phase, which is when the findings in McAllister’s studies were obtained. One cannot generalize findings from the acute or subacute post-injury phase to the chronic post-injury phase. Another over-generalization is that the 19991 study only contained 11 subjects and the 2001 study included 18 subjects. These are two few patients to generalize to broad patients populations.
Another problem is that the above argument assumes that fMRI activation data is related to fatigue or other changes in real-world functioning in these patients, when the studies did not demonstrate this. The argument also assumes that neuropsychological evaluation results, which have been established for decades in the scientific literature as the gold standard for assessing brain functioning should be superseded by a technique (i.e., fMRI) that is mostly constrained to the research setting at this point. Along these same lines, it should be noted that none of the patients in McAllister’s studies showed any evidence of brain damage on structural neuroimaging scans.
The last, and perhaps, most important problem with the inferences that have been made from McAllister’s study is the effect that the visual image from the fMRI data can potentially have upon patients and health care providers when not provided in the proper context. That is, patients seeing such an image may form the incorrect belief that this is what must be happening to their brain, leading them to believe that their symptoms (or those of a loved one) are all neurologically based when in reality, there may be multiple non-neurological explanations for symptom persistence. Health care providers who do not read the original studies may also form incorrect impressions based on this information, leading to incorrect diagnosis and treatment.
REFERENCES
McAllister, T.W. at al. (1999). Brain activation during working memory 1 month after mild traumatic brain injury: A functional MRI study. Neurology, 53, 1300–1308.
McAllister, T.W. at al. (2001). Differential working memory effects after mild traumatic brain injury. Neuroimage, 14, 1004–10012.
Dr. Carone offers paid lectures on MTBI, “post concussion syndrome,” and symptom validity testing upon request. He can be contacted at info@mtbifacts.com.
Copyright © 2009, MTBIFacts.com. All rights reserved. You may not reprint or redisplay this material for commercial use without the express written consent of MTBIFacts.com.
Use of MTBIFacts.com is subject to reading this disclaimer.
|
