Saturday, May 15, 2004

Recognition of Facial Emotions in
Neuropsychiatric Disorders

The second article in The April 2004 issue of CNS Spectrums is titled Recognition of Facial Emotions in Neuropsychiatric Disorders.  It was written by Christian G. Kohler, MD, Travis H. Turner, BS, Raquel E. Gur, MD, PhD, and Ruben C. Gur, PhD.  Most of them are from the University of Pennsylvania.  I'm not sure why this article is in with the others, since it is not really about neuroimaging.  The abstract is here:

Recognition of facial emotions represents an important
aspect of interpersonal communication and is governed
by select neural substrates. We present data on emotion
recognition in healthy young adults utilizing a novel set
of color photographs of evoked universal emotions.
In addition, we review the recent literature on emotion
recognition in psychiatric and neurologic disorders, and
studies that compare different disorders.
CNS Spectr. 2004;9(4):267-274

The authors claim that certain facial expression of emotion are universal; others are subject to variation from one culture to the next.  The universal ones are: happiness, anger, fear, sadness, and disgust.  Variable expressions include shame, arrogance, contempt, guilt, flirtatiousness, and admiration.  [Editorial note: travelers are advised to remember that flirting is not  recognized universally.]  They report that the classical hypothesis is that much of the processing that our brains go through to interpret facial expressions takes place in the right hemisphere.  Some researchers believe that the right hemisphere is best at interpreting all kinds of emotional expressions.  Others feel that the left hemisphere is used more in the interpretation of positive feelings; the right being better at processing negative feelings.  The authors report that they conducted a meta-analysis of 65 pertinent neuroimaging studies, and that support for these notions of lateralization is limited, but not absent.

The reason that the processing of facial cues of emotional expression is important is that problems in such processing occur in certain types of brain injury, developmental disorders, and in thought disorders, especially schizophrenia.  In persons with no injury, developmental disorder, and with no thought disorder, the recognition of facial emotional expression is an important part of communication.  

They report on the various methodologies used, and the limitations of earlier methods.  Specifically, the use of greyscale versus color images, and the use of images of only Caucasians, introduce problems.  They worked with a set of 96 images rendered in color, and in 3D, to try to overcome some of the limitations of earlier studies. 

One of their earliest findings was that the amygdala is activated in tasks of recognition of facial emotional expressions, and that this activation is muted in persons with schizophrenia.  They go on to report on the accuracy of recognition in normal subjects, then review various pathological states, including schizophrenia, mood disorders, brian injury, neurodegenerative diseases, autism, and mental retardation.

Among normal subjects, happiness is the expression that is recognized most precisely, followed by fear, sadness, anger, and disgust.  With all of these emotions except disgust, error rates are lower when the pictures illustrate emotions expressed with greater intensity. 

With regard to schizophrenia, they point out that impairment of the capacity to identify accurately the various expression of emotion is mentioned in even the earliest case reports.  It is considered to be a significant source of social impairment.  This appears to be true, independent of the effects of other kinds of symptoms (hallucinations, delusions, and cognitive impairments.)   They report that the difficulty in persons with schizophrenia is at least partly localized to the mesial temporal regions of the brain.  The difficulty occurs both with regard to correct identification of emotional expression, and with regard to incorrect attribute of emotion when the stimulus is a picture of a neutral expression.  Also, persons with schizophrenia did not achieve greater precision when the intensity of the expressed emotion increased.  the authors do not speculate as to why this latter finding would occur.  However, it has been recognized for a long time that persons with schizophrenia are more prone to have problems with symptoms when they are in the presence of others who are engaging in intense expressions of emotion.  It is as though the more intense expressed feelings somehow provoke greater interference with cognitive processing.  It is tempting to hypothesize that the emotional interference with cognition is responsible for their finding, that persons with schizophrenia do not attain greater precision when asked to identify stronger expression of emotion.  I would like to see a study that uses functional neuroimaging to test this hypothesis.

With regard to bipolar disorder, the authors report on studies with stable bipolar patients, meaning that they were neither manic nor depressed at the time of the study.  These persons had difficulty accurately rating the intensity, but not the quality, of emotional expressions.  By this they mean that patients with bipolar disorder can accurately identify what kind of emotion they are seeing, but not how intense the expression is.  In contrast, persons who are actively manic at the time of the study have greater difficulty recognizing negative expressions.  This is consistent with other observations, that persons who are manic tend to see everything in a positive light.

Turning their attention to major depression,  they note that there have not been very many studies.  Perhaps this is because the studies that have been done so far are not very interesting.  At least one study showed a negative bias, indicating that people who are depressed are more likely to view others as being depressed. 

Next to be considered was the topic of developmental disorders.  This would include mental retardation and autism.  There is a bit of a controversy about whether persons with mental retardation have a separate impairment of their ability to recognize emotions, or whether their difficulties are due to a general cognitive impairment.  Review of the studies shows scant evidence for a selective impairment.  Even among those with autism, there is no particular evidence for a selective impairment.  Any difficulties experienced by persons with autism were correlated with general intelligence. 

The authors then considered the effect of brain lesions.  The start by mentioning Urbach-Wiethe disease.  (Yeah, I had to look it up, too. Link-free registration required)  This is a disease that, among other things, causes selective destruction of the amydgala.  As expected, such persons have impaired ability to recognize expressions of fear; sometimes, recognition of other negative emotions is impaired as well.  Studies on persons who have had a stroke indicate that a loss of right hemisphere tissue have greater impairment of recognition of facial expression of emotion than those with lesions on the left. 

Alzheimer's Disease tends to cause early loss of mesial temporal lobe structures, including the hippocampus and amydgala.  This leads to impairment of recognition of facial expression of emotion, as expected.  Persons with frontotemporal dementia tend to have even greater impairment.  However, the frontotemporal dementia patients did have sparing of the ability to recognize fear.  

Return to the Corpus Callosum, here.

(Note: The Rest of the Story/Corpus Callosum has moved. Visit the new site here.)
E-mail a link that points to this post:
Comments (0)

Thursday, May 13, 2004

Functional Neuroimaging:
Introduction and First Article

The April 2004 issue of CNS Spectrums is titled Neuroimaging of Emotions in Psychiatry.  The forward was written by the guest editor,  Israel Liberzon.  Dr. Liberzon is associate professor of psychiatry and co-director of the Trauma, Stress, and Anxiety Research Center in the Department of Psychiatry at the University of Michigan Medical School.  Much of his work was done here.

This post includes a discussion of the Introduction to the CNS Spectrums issue and a review of the first article.  The article itself describes some of the corellations found between activation of specific parts of the brain, and general vs. specific emotional tasks.  I include some comments about the potential practical significance of this kind of work.  Please refer to my previous posts (1  2) for an introduction to the topic.  For a quick review of neuroanatomy and neuroimaging, go here

Some of you may have heard an interview with Dr. Liberzon.  He was interviewed on the Todd Mundt  show on Sept. 18, 2001.  Understandably, Posttraumatic Stress Disorder was an item of interest to the media at the time.  Unfortunately, the streaming audio is not available any longer, although you can order a tape if you want.  In any case, Dr. Liberzon is a bit of a character, and a very good scientist who is interested in the neurobiology of PTSD.  A quick (not necessarily exhaustive) Medline search turns up 26 articles that he has published in his career.  An example is Brain-Imaging Studies of Posttraumatic Stress Disorder.

In his introduction (which in not online), he points out that functional neuroimaging has "revolutionized the field of cognitive neuroscience," adding that "Investigators are now able to identify the neurocircuitry associated with specific cognitive functions."  What he means is that it is possible to present a person with a cognitive task and see what specific parts of the brain become more (or less) active when the task is performed.  Equally important, it is possible to do the same thing with emotional stimuli.  That is, you can present the person with a stimulus to provoke a particular emotional state, and see what parts of the brain get triggered.  He states, "This work carries a great promise to facilitate our our efforts to advance understanding of abnormal emotional functioning in psychiatric illness." 

From time to time, one hears that psychiatry is "not really a science," or "it's all intuitive."  those who may hold such beliefs may be interested to read what Dr. Liberzon points out: "This [neuroimaging] approach also allows us to examine empirically prevailing but difficult to test beliefs or hypotheses, such as the predominance of right hemisphere in processing emotions."  Regarding one of the papers, he comments: "They demonstrate how major hypotheses, such as the involvement of serotonergic systems in affective disorders, can be critically examined in vivo, and how a new hypothesis regarding the involvement of opioidergic systems, can be derived from neuroimaging findings. 

Science, of course, is all about generating and testing hypotheses. 

The editor of CNS Spectrums is Dr. Jack Gorman.  In the forward to the issue, he begins, "In most medical specialties, detailed knowledge of the normal anatomy, physiology, and histology of the organs of interest were known before the specifics of disease states were revealed."  He adds, "In psychiatry, and to some extent neurology, we have tried to do things in the opposite direction."  There is some precedent fro this.  The field of speech pathology was in a state of prolonged infancy until World War II.  Then, studies on soldiers with head and neck injuries led to remarkable advances in the field. 

In some ways, the strategy of studying disease states to understand normal function Is like the modern developments in some technical fields.  For example, the space program led to the development of all kinds of technology that nowadays is used in everyday living.  The principle here is that you study the exceptions, or extreme cases, in order to further your understanding of ordinary things. 

The first paper in the issue is Functional Neuroimaging Studies of Human Emotions, By K. Luan Phan, MD, Tor D. Wager, PhD, Stephan F. Taylor, MD, and Israel Liberzon, MD.  In addition to Dr. Liberzon, the only one I know is Dr. Taylor.  Dr. Taylor is one of those quietly brilliant people that inspires awe in ordinary intelligent people.  Dr. Phan trained at the  University of Michigan, but  I don't recall meeting him.  He started his training about the time I finished.  The paper they wrote is a survey of 55 PET and fMRI studies that were done to assess various aspects of brain function in response to emotional stimuli.  They set out to basically develop an atlas of corellations between different emotional responses and different parts of the brain.  This is something that is inherently difficult to do.  As you might expect, different brains can have widely different responses to the same stimuli.  In order to try to make sense of this, it is necessary to study several different people and average the results.  One of the strongest corellations they found was the relationship between fear and increased metabolic activity in the amygdala.  (If you are curious to know where all the different brain parts are, see this guide in Wikipedia.  You can see pictures of brain dissection here, and a neuroimaging atlas of the human brain here.) 

From the Harvard Whole Brain AtlasAnother strong correlation they found was that the medial prefrontal cortex often is activated in response to generic emotional stimuli.  By this I mean that the activation occurred without regard to the type of emotion involved.  As they put it: "The findings suggest that the MPFC may have a 'general' role in emotional processing."  Although they found that "no single brain region is commonly activated by all emotional tasks," the MPFC was the one that was turned on most often.  Its location in the prefrontal cortex suggests that it may be a linkage point between emotional processing and cognitive processing: one of the bridges between thinking and feeling.  They point out that "several studies have been recently published demonstrating that when subjects turn their attention inward toward themselves, as often required during general emotional processing, activity within the MPFC is increased.  Studies requiring subjects to determine if personality trait adjectives are descriptive of themselves (versus someone else) or to reflect on their own abilities/traits/attitudes have observed engagement of the MPFC."

Thus, all those internet quizzes that we see from time to time actually exercise your MPFC. 

Another area of interest in emotional neuroscience is the anterior cingulate cortex.  It has been know for a long time that injury to the ACC can lead to wither apathy or lability (instability).  Therefore, it is tempting to speculate that the ACC is involved in the regulation or modulation of emotional responses.  Their review of neuroimaging studies of the ACC shows that it is involved in detecting emotional cues.  For example, when watching a movie, your brain is flooded with all kinds of stimuli, but only some of them have any emotional significance.  The ACC seems to be involved in distinguishing the emotional figure  from the background noise.  The authors speculate about some possible functions: "As a detector of salient information in general, the ACC could serve to allocate brain resources, heighten sensitivity and direct attention to environmental cues produced by the evocative stimulus."  They found also that the ACC is activated when people are asked to recall events that are linked to a particular emotional state. 

Another finding with regard to the ACC is that it often is activated when people feel sad.  In fact, activity there is increased during the depressive phase of some types of major depressive disorder.  This kind of finding is interesting because it indicates that neuroimaging studies may someday be useful in the task of making diagnoses that are more specific than those made using current descriptive methods.  This, in turn, could lead to better selection of treatments. 

The third and final brain region described in the article is the insula.  The insula, like the ACC, is activated during emotional recall tasks.  Unlike the ACC, the insula is not activated by viewing external cues, such as movies, with emotional content.   This finding has been duplicated in studies of non-human primates.  It appears that the insula is activated most strongly in response to aversive or threat-related emotions.  Some, but no all, studies showed a specific activation of the insula in response to disgust.  It is activated when the brain receives visceral (primarily gastrointestinal) aversive sensations.  the authors speculate that the insula may be the center that is responsible for literal "gut feelings." 

The authors conclude that there is emerging evidence that some brain structures are involved in specific emotional states, whereas others serve a more general function of emotional processing.  They are hopeful that these and future studies will permit the understanding of the functional neuroanatomy of emotion. 

I already have commented that this kind of work shows that psychiatry does have a foundation in hard science, and that the generation and testing of hypotheses is possible using neuroimaging studies.  The paper by Phan et. al. shows that we are beginning to find ways of possibly increasing the specificity of psychiatric diagnoses.  This may enable us to select treatments ahead of time that are more likely to work, as opposed to the current process of trying one thing after another, without much specific indication of what treatment is most likely to do the job.

Return to the Corpus Callosum, here.

(Note: The Rest of the Story/Corpus Callosum has moved. Visit the new site here.)
E-mail a link that points to this post:
Comments (0)

Wednesday, May 12, 2004

fMRI in the News

...Well, not exactly.  CNS Spectrums is not what you would call a leading news outlet.  Still, it devoted an entire issue to the subject; I just got it today.  Functional Magnetic Resonance Imaging of the human brain is one of those obscure things that gets mentioned in the news occasionally, but it is not one of those things that reports pick up on and report reflexively.  Probably, that's because most of the individual research findings are so obscure as to be meaningless to someone who does not have a broad understanding of the scientific or clinical context of the discovery.

The functional  part of fMRI refers to the fact that the technique permits imaging of the function of an anatomical part, as opposed to just getting a picture of the part itself.  With fMRI, it is possible to see what parts of the brain become more (or less) active when certain tasks are performed. 

fMRI is mentioned infrequently in the Blogosphere.  According to Waypath, it is mentioned, on average, less than once per day.  Most of these are cursory mentions.  There was a bit of interest in an fMRI study showing the anatomic location of the placebo effect.  (Scientific American, February 20, 2004, Scientists See How Placebo Effect Eases Pain), and in an article about using fMRI to detect differences in the way the brains of Democrats and Republicans viewed emotionally-charged political images.  Aziz Poonawalla, on the blog UNMEDIA wrote  about the Democrat/Republican study as reported in the NYT.  It also was mentioned on JawsBlog  and Drudge Report.  Of greater clinical interest, there was a report showing possible utility of real-time biofeedback using fMRI to teach people to control pain sensations.  This was mentioned on Marginal RevolutionPain for Philosophers, and  Seedlings & SproutsDean's World  makes a brief mention of a study  on the localization of long-term memory.  This was not a study that used fMRI, but Dean's comments about the fact that there must be a limit to the amount of information the brain can hold.  This is something that has been investigated using fMRI.  FuturePundit elaborates on this, and adds some comments about a different study  that seems to show what happens in the brain when a person has an eureka moment.  The latter study is especially interesting because the authors show a correlation between the findings using fMRI and those revealed using EEG. 

fMRI is playing a key role in the Human Brain Project.  The HBP is a project, comparable in score to the Human Genome Project, that attempts to use bioinformatics technology to develop a complete picture of human brain function.  As Dean mentioned, there is a limit to how much information the brain can store.  This, ironically, limits out ability to understand the brain.  The HBP is an attempt to address this limitation.  As they say in their introduction:

Understanding brain function requires the integration of information from the level of the gene to the level of behavior. At each of these many and diverse levels there has been an explosion of information, with a concomitant specialization of scientists. The price of this progress and specialization is that it is becoming virtually impossible for any individual researcher to maintain an integrated view of the brain and to relate his or her narrow findings to this whole cloth. Although the amount of information to be integrated far exceeds human limitations, solutions to this problem are available from the advanced technologies of computer and information sciences.

In the first paragraph of this  post, I mentioned that the results of individual studies often are meaningless without a broad understanding of the scientific context.  This applies to scientists as well as laypersons.  The HBP may be helpful in helping specialists have ready access to the information required to place a given study in a meaningful context. 

By the way, for those regular folks who want to develop an understanding of the scientific context of various studies, a good way to find recent articles on scientific topics is to do a keyword search at ScienceDaily.com.  Here  is the result of a search for fMRI.  This does not give you the complete context, but it is a good place to start.

fMRI is not the only tool that allows for functional imaging of the brain.  Positron Emission Tomography and Single Photon Emission Computed Tomography have been in use for years.  In general, though, fMRI is the most practical of the currently-available methods, since it does not require the injection of radioactive tracers.  A related technique, magnetoencephalography, is under development.  Furthermore, MRI technology is constantly being improved

All of this is really background information for what I plan to post next, which is a review of the CNS Spectrums issue devoted to neuroimaging.     

Go back to The Corpus Callosum here.

(Note: The Rest of the Story/Corpus Callosum has moved. Visit the new site here.)
E-mail a link that points to this post:
Comments (0)

Monday, May 10, 2004


First, first out if it works.

(Note: The Rest of the Story/Corpus Callosum has moved. Visit the new site here.)
E-mail a link that points to this post:
Comments (0)