Brain Imaging and Conduct Disorder: Temporal Lobe Abnormalities

Conduct disorder is a complex behavioral disorder with significant risk for later adult psychopathology.

There is increasing evidence for a biological basis for conduct disorder.

Twin studies show a significant genetic contribution to the disorder.

Brain imaging studies also point to biological factors in conduct disorder.

Gregory Wallace and colleagues recently published a structural MRI study of conduct disorder in 22 adolescents between the ages of 10 and 18. Conduct disorder subjects were compared to a group of 27 age-matched controls on imaging measures.

This study focused on measures of brain cortex thickness, brain surface area and degree of brain folding or gyrus formation.

Conduct disorder was linked to the following brain structural abnormalities:

• Reduced cortical thickness in the superior temporal lobes
• Reduced gyrus formation in the ventromedial frontal cortex
• Reduced volume of the the amygdala and striatum (putamen and pallidum) 

The research group also found a negative correlation between superior temporal lobe thickness and psychometric measures of callousness/unemotional styles.

The mechanism for the temporal lobe to be involved in the symptoms of conduct disorder is unclear. Cortical thinning in this region has been found in adults with psychopathy. 

The authors note that amygdala/temporal lobe integration is necessary for stimulus-reinforcement learning. This integration may explain some of the deficits found in the current study.

This study will be important for continuing research in the genetics and pathophysiology involved in conduct disorder. Intervention strategies will need to address potential biological deficits contributing to the behavioral and learning problems in conduct disorder.

Readers with more interest in this research can access the free full-text manuscript by clicking on the PMID link in the citation below.

Image of brain with superior temporal lobe highlighted in green is an iPad screen shot from the Brain Tutor app.

Follow the author on Twitter @WRY999

Wallace GL, White SF, Robustelli B, Sinclair S, Hwang S, Martin A, & Blair RJ (2014). Cortical and subcortical abnormalities in youths with conduct disorder and elevated callous-unemotional traits. Journal of the American Academy of Child and Adolescent Psychiatry, 53 (4), 456-650 PMID: 24655655

Fatherhood and the Biology of Infant Care Behavior

Early positive parental-infant attachment provides a key developmental advantage.

The maternal-infant dyad has received the most research attention. However, fathers also can play an important role in infant development. An increasing number of studies focus on the biology of fatherhood and infant care behavior.

Ilanit Gordon along with colleagues in Israel and the United States examined the role of two central neuropeptides in paternal behavior during the first six months of infancy.

In their study, 43 fathers with a firstborn infant were studied when the infants were two and six months old. Fathers were rated on levels of emotional synchrony with their infant. Additionally, fathers were rated on levels of paternal involvement in exploratory play.

Paternal serum oxytocin and prolactin levels were obtained at both the two and six month time points. Oxytocin has been identified as a key neuropeptide in maternal-infant bonding, but it's role in fathers is unclear.

The authors found significant relationships between the two hormones and paternal-infant behavior. Paternal serum oxytocin levels correlated with emotional synchrony while serum prolactin levels correlated with coordinated exploratory play.

The authors note in their discussion that prolactin had previously been identified as important in animal studies of fathers but not in humans.

The mechanism of how these two hormones impact paternal behavior is unclear.

A recent study by Mascaro and colleagues suggest neuropeptides may influence specific brain regions known to be important in empathy (anterior insula and inferior frontal gyrus).

In their study, a series of fathers were studied for responsiveness to the cries of children between one and two years of age. Brain regional activation levels were compared across a paternal responsiveness to infant cries.

The key finding in this study was a nonlinear relationship between anterior insula activation and paternal responsiveness. Fathers with a moderate level of activation outperformed fathers with low or high levels of activation. 

These types of studies are important. They may hold promise for improving parental attachment to their children. Additionally, they may provide some insight into the biology of child abuse and neglect. These first studies are limited correlational design studies but they add evidence of the need for additional more complex and sophisticated research.

Readers with more interest in these two studies can get more information by clicking on the PMID links in the citations below.

Image of brain amygdala is a screen shot from the iPad app 3D Brain.

Follow the author on Twitter @WRY999

Gordon I, Zagoory-Sharon O, Leckman JF, & Feldman R (2010). Prolactin, Oxytocin, and the development of paternal behavior across the first six months of fatherhood. Hormones and behavior, 58 (3), 513-8 PMID: 20399783

Mascaro JS, Hackett PD, Gouzoules H, Lori A, & Rilling JK (2014). Behavioral and genetic correlates of the neural response to infant crying among human fathers. Social cognitive and affective neuroscience, 9 (11), 1704-12 PMID: 24336349

Real-Time fMRI Psychotherapy

Old World Psychotherapy: Sofa of Sigmund Freud

There has been a series of interesting research studies examining the effect of psychotherapy on brain structure and function.  These studies have typically shown that effective psychotherapy results in reduction of brain deficits or abnormalities associated with a specific neuropsychiatric disorder.

Now a study published in Plos One summarizes the results of study examining the use of real-time fMRI to provide neurofeedback during an amygdala activation task.

This research was completed by neuroscientists affiliated with the Laureate Institute of Brain Research in Tulsa, Oklahoma and George Mason University in Fairfax, Virginia. (Disclosure: The author of Brain Posts is employed by Laureate Institute of Brain Research but was not involved in the study reviewed in this post.)

The authors of this study noted the key role of the amygdala in the processing of emotions.   They developed a experimental paradigm to train control subjects to increase the activation of the brain left amygdala.  A group of young male subjects were instructed in a happy autobiographical memory task and provided real-time feedback on how successful they were in increasing blood flow to the left amygdala.

Subjects identified three key happy memories from their past.  During the experimental phase, they were instructed to recall these specific memories while being scanned using an fMRI scanner. They were provided real-time feedback on a monitor screen on the changes in left amygdala BOLD signal.  (Subjects were told prior to scanning that fMRI neurofeedback is delayed by a few seconds due to the brain hemodynamic process).

Subjects provided real-time feedback were more successful at increasing the left amygdala activation than those in a control group.  This increase in the experimental group correlated with increases in other brain areas known to have functional connectivity with the amygdala (fronto-temporo-limbic network).

New World Psychotherapy: Real Time fMRI

Additionally, the study identified six specific regions where functional connectivity identified correlations with the left amygdala activation:

• right medial frontal cortex
• bilateral dorsomedial prefrontal cortex
• left anterior cingulate cortex
• bilateral superior frontal gyrus

Subjects were selected based on being free of a history of neuropsychiatric disorders including anxiety and depression.  However, there was some variability in the level of change in left amygdala activation with neurofeedback training.  Subjects who scored high on the Difficulty Identifying Feelings scale had less increase in the left amygdala.  Additionally, subjects with higher scores on a scale of being susceptible to anger showed less increase.

This research is an very important advance in understanding the amygdala and regions connected with the amygdala.  Additionally, it raises the possibility that real-time fMRI may emerge as a tool to understand processes associated with psychotherapy and to be an emerging model for providing therapy under real-time neurofeedback conditions.  

The site of psychotherapy might be moving from the sofa model of Sigmund Freud to the fMRI scanner.  Both methods have subjects that lie down, but only the fMRI method provides real-time feedback of brain effects related to a psychotherapy intervention.

The authors note that this study was a type of "proof-of-concept" study since it focused on healthy control subjects.  They suggest that this type of model might be particularly relevant to cognitive behavioral treatment of conditions such as PTSD and major depression.

Photo of Sigmund Freud sofa from the Freud Museum in London from Wikipedia distributed under the GNU Free Documentation License.


Photo of Functional Magnetic Research Imaging device courtesy of the Laureate Institute for Brain Research. 

Zotev, V., Krueger, F., Phillips, R., Alvarez, R., Simmons, W., Bellgowan, P., Drevets, W., & Bodurka, J. (2011). Self-Regulation of Amygdala Activation Using Real-Time fMRI Neurofeedback PLoS ONE, 6 (9) DOI: 10.1371/journal.pone.0024522

Autism and MRI Physical Biomarkers

Minor physical anomalies (MPAs) commonly occur in those with autism.  I have previously published a post on a study outlining the type and prevalence of these anomalies in a series of case of autism and austim spectrum disorder.

One of the MPAs noted in the 1970s in autism spectrum disorder is an increased intraorbital distance (distance between the eyes).   This abnormality also noted as hypertelorism has been noted in a variety of brain developmental abnormalities as well as in some normal individuals.

The brain developmental correlates of hypertelorism have not been studied extensively.  It is possible that developmental hypertelorism may reflect brain developmental variations linked to clinical disorders.

Cheung et al and colleagues from the University of Hong Kong and Harvard School of Dental Medicine recently published a brain MRI study of intraorbital distance in autism spectrum disorder in PloS One.  This study examined the correlation of intraorbital distance with a variety of brain structural measurements.

Thirty six children between the ages of 7 and 16 years with autism spectrum disorder were compared to a group of 55 developmentally normal children matched by age and gender.  The two groups were in the normal intelligence range with the verbal IQ of the autism spectrum group 112 compared to a verbal IQ of 117 in the control group.

Magnetic resonance imaging scans were used to accurately measure the intraorbital distance.  This measurement was then compared with brain structure volumes.

Intraorbital distance correlated with several brain regions volumes in the autism spectrum group but not in the control group.  The areas with increased volume correlating with intraorbital distance in the autism spectrum group included:

• Bilateral amygdala
• Bilateral medial temporal lobe regions
• Left inferior frontal cortex lobes

Amygdala developmental abnormalities have previously been noted in studies of autism.  The medial lobe appears to be important in social and language skills developments.

The authors note the possible relationship between intraorbital distance and brain development in this statement from the manuscript discussion section: 

     "The inference is that in this group, the growth of midline bony and brain regions are tightly linked; that is, regions involved in the regulation of socialization, emotion and memory appear to enlarge with the visual system".

These finding suggest that MRI intraorbital distance may be a potential biomarker for autism spectrum disorders.   Additional longitudinal studies of intraorbital distance and brain development in children may provide additional support for the findings in this cross-sectional studies.

Image of intraorbital distance measurement in autism from Cheung et al distributed under the terms of Creative Commons Attributions License which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. 

Cheung C, McAlonan GM, Fung YY, Fung G, Yu KK, Tai KS, Sham PC, & Chua SE (2011). MRI study of minor physical anomaly in childhood autism implicates aberrant neurodevelopment in infancy. PloS one, 6 (6) PMID: 21687660

Brain Basis for Emotion Recognition Deficits in Depression

There is a emerging understanding of the role of social perception problems in depression and anxiety disorders.  Depression appears to effect the cognitive ability to judge the facial expression of others.  This impairment poses a challenge for interpersonal function and social relationships.  Research is now pinning down the neural basis for this deficit and to determine it’s persistence and the effect of depression remission on this social cognition function.

van Wingen and colleagues from the Netherlands recently published an fMRI study on this topic in Psychological Medicine.  The study had the following elements in research design:

• Subjects: Twenty case subjects with first episode of major depression (medication naive), Twenty one case subjects recovered from a first episode of depression and 30 healthy individuals without a history of depression
• f MRI Task: Visual recognition of anger or fearful face by semantic labelling or visual matching compared to a control task of matching facial orientation without attention to emotion
• Additional Neuropsychological Testing: Depression symptom level, anxiety (state and trait) symptom level, IQ, memory, visual learning, attention, psychomotor speed and executive function.

The depression case subjects were not different from recovered depressed subjects and controls on most of the tests of neuropsychological function.  As expected, they did have higher depression symptom severity scores (Hamilton Depression Scale average for the depressed group was 21.8) and higher anxiety symptom severity scores.

The depressed group performed as well as the remitted group on the control task and the visual emotion matching task.  However, they performed worse than both groups on the semantic matching task, i.e. selecting the correct word label for the emotion displayed visually.  

The brain regions that correlated (increased activation) with impaired semantic emotion labeling included three distinct regions:

• right amygdala
• left inferior frontal gyrus
• anterior cingulate cortex

The authors propose that one explanation for this finding in three distinct brain regions.

“The left inferior frontal gyrus is thought to integrate language with other information (Hagoort, 2005;Willems et al. 2007). Therefore, we suggest that the inferior frontal gyrus may integrate the semantic knowledge about the concepts of anger and fear with the emotional information conveyed by the faces by interacting with the amygdala. The concurrent activation of these systems may subsequently trigger automatic negative thoughts and stimulate task unrelated processes such as rumination (Siegle et al. 2002; Ray et al. 2005), and thereby hinder task appropriate behaviour.”

This explanation fits with the cognitive behavioral theory of depression.  The core concept being that depression is characterized by increased automatic negative thoughts about the self, the future and the environment.  The thoughts are generated by a maladaptive negative cognitive schema.  Activation of these automatic thoughts (i.e. by being shown negative facial emotions) produces distraction from tasks (i.e. correctly labelling these facial emotions). This process appears limited to when depression is present as the recovered depression group showed no deficit.

The authors also conclude the findings could be due to a compensatory mechanism for inadequate behavior or an altered coping mechanism for dealing with demanding situations. 

This study provides additional support that fMRI research may lead to advances in the treatment of depression by both psychotherapy and psychopharmacologic interventions.

Screen shot of 3D Brain showing the limbic system structures amygdala and cingulate cortex (along with the inferior frontal cortex) thought to be involved in impaired facial recogniton in depression.

van Wingen, G., van Eijndhoven, P., Tendolkar, I., Buitelaar, J., Verkes, R., & Fernández, G. (2010). Neural basis of emotion recognition deficits in first-episode major depression Psychological Medicine, 41 (07), 1397-1405 DOI: 10.1017/S0033291710002084

Neuroethics: The Brain and Political Beliefs

Article first published as Brain Science and Political Belief on Technorati.


Brain science is providing some important insights into the mechanisms involved in a variety of beliefs including political, religious and moral beliefs. Dr. Jordan Grafman, Ph.D. currently with the Kessler Foundation has led some of the key research initiatives in this area of brain research. He recently presented at the May, 2011 Warren Frontiers in Neuroscience lecture series in Tulsa, Oklahoma. I previously posted a summary of work relevant to the brain science and moral beliefs and a separate post on religious beliefs. Here are some of the key points from his presentation related to political beliefs and relevant published research manuscripts.

A important starting point in understanding how the brain processes political beliefs is to discover key elements of these beliefs. The classical description of political beliefs defines individuals long a single criterion domain, conservative to liberal. But statistical modeling of large numbers of individuals demonstrates three key domains for political belief. These domains appear somewhat independent of each other and appear to engage different brain neural circuits. The three domains and localized areas the brain involved include:

• individualism--medial prefrontal cortex and the temperoparietal junction
• conservatism-dorsolateral prefrontal cortex
• radicalism-ventral striatum and posterior cingulate

So how does brain circuitry differ between those who are interested in politics compared to those with little interest in the area? Grafman while at his previous NIH position and colleagues from Italy as well as George Mason University in the U.S. examined this question in a series of 25 subjects using functional magnetic imaging scanning technology.

Subjects in the scanner were asked to agree or disagree with a variety of political opinions. Subjects who were interested in politics showed significantly more activation of the brains regions in the amygdala and ventral striatum. Subjects disinterested in politics showed limited activations in these regions regardless of the content of religious beliefs encountered.

The amygdala has been demonstrated to be an important region for processing both positive and negative emotional stimuli. The ventral striatum appears to be a key region in processing reward and positive affect.

The research suggests those with strong political interests, i.e. "political junkies" engage brain circuits involved in emotional and reward reinforcement. Similar circuits appear to be engaged in other individuals by chemical as well and specific environmental stimuli. For example, similar circuits may be activated by religious stimuli in those with strong religious beliefs.

It is unknown exactly how the brain influences each individuals selection of an area of interest and engagement. We do not know why one individual is drawn to a field of technical interest, i.e. computer science rather than a field of social interest such as politics.

But once an individual establishes strong political interest and a strong political belief system brain circuitry effects develop. These circuits appear to not be specific to politics but specific to circuits controlling emotional engagement and reward reinforcement.

Future brain science research is likely to further discover the mechanisms involved in political belief development and maintenance. An ethical challenge is likely to emerge on how these discoveries are used by political parties to influence voter preferences and voter engagement.

Link to Dr. Grafman's website related to his research on traumatic brain injury at the Kessler Foundation.

Zamboni G, Gozzi M, Krueger F, Duhamel JR, Sirigu A, & Grafman J (2009). Individualism, conservatism, and radicalism as criteria for processing political beliefs: a parametric fMRI study. Social neuroscience, 4 (5), 367-83 PMID: 19562629


Gozzi M, Zamboni G, Krueger F, & Grafman J (2010). Interest in politics modulates neural activity in the amygdala and ventral striatum. Human brain mapping, 31 (11), 1763-71 PMID: 20162603


Photo of African Elephant, the symbol of U.S. Republican Party from author's private family collection by photographer Sarah Yates.

Anxiety as a Gut Feeling: Understanding Interoception

Marcus Paulus presented the April 2011 Warren Neuroscience Frontiers in Neuroscience Lecture.  The presentation was titled: Interoception and Anxiety.

Interoception is the summation of a variety of bodily perceptions that make up the integrated sense of our own physiological state.  Perceptions included in interoception include: pain, temperature, tickle, sensual touch, stomach discomfort to due acidity, air hunger and muscle tension.  Here are my notes from Dr. Paulus' presentation and his research manuscript on this topic area.

• Anxiety proneness is a trait that can be measured and is associated with high risk of later development of an anxiety disorder
• Anxiety proneness linked to increased activation of the dorsal amygdala and the anterior insula in brain fMRI tasks such as the Emotion Face Assessment task of Hariri
• Patients with anxiety also show insular hyperactivation in anticipation of negative cues
• Benzodiazepines like Valium reduce activation of the insula as well as the amygdala in response to angry faces

• There is growing awareness the brain insular cortex plays a key role in interoception--receiving signals from the body and integrating these signals with emotional response and regulation (see a previous post summarizing the function of the insula and possible roles in clinical neuroscience disorders)
• The insula also connects to a central pathway important in anxiety involving the anterior cingulate cortex and the dorsolateral prefrontal cortex--these areas provide input to the insula for planning and acting in the face of

• Key properties of the interoception include the signals from internal organs including the lungs, heart, gastrointestinal tract and genitourinary systems
• Many of these signals provide awareness of body and help promote homeostasis
• These signals also are involved in our sense of self and the passing of time

• A new area of understanding is the important role of personal beliefs in emotional processing--personal beliefs may modulate interoception and the perception of emotional cues
• A belief that a situation or cue is more dangerous than it really is, i.e. I will embarrass myself at the party, can modulate how emotion is processed, and can amplify a anxious response to the situation

Future research in the area of interoception and anxiety will target:

• Genetic influences on interoception
• How cognitive interventions may influence dysfunctional beliefs related to anxiety
• How interoception may help with more biological classification of types of anxiety
• Can people be trained to up or down regulate the insular cortex to reduce anxiety?
• How treatments for anxiety effect the elements of interoception

Brain Tutor iPad Screenshot of Insular Cortex in Green Courtesy of Author

Paulus MP, & Stein MB (2010). Interoception in anxiety and depression. Brain structure & function, 214 (5-6), 451-63 PMID: 20490545

Emotional Processing Bias in Depression

Clinicians and individuals with depression understand the tendency for depression to be associated with over-interpretation of negative cues in the environment.  Depression seems to heighten perception of negative environmental cues including interpersonal (or social) cues.   The cognitive behavioral model of depression emphasizes the cognitive triad—a negative bias (view) of the self, the environment and the future.

Functional magnetic resonance imaging is providing a model to study emotional processing and better understand how this processing may be disturbed in depression.  When shown brief images of emotion-laden faces, subjects suffering from depression show exaggerated responses.  Depressed or angry faces produce heighted amygdala responses in those with depression.  Happy faces produce a blunted amygdala response. 

Teresa Victor and colleagues recently published a study in the American Journal of Psychiatry providing additional information about this emotional processing bias.  (Disclosure:  Dr. Victor is now a neuroscientist with the Laureate Brain Institute-my employer).   She summarized her findings in a recent journal club.  Here are the key findings from her recent research:

• The amygdala response to facial images occurs even with brief subliminal (unconscious) presentations
• Sad facial images activate the amygdala in depressed subjects (compared to controls) in both those with active and remitted depression
• Happy facial images activate the amygdala in controls more than depressed subjects
• Eight weeks of selective serotonin reuptake inhibitor therapy (sertraline) reverse  (normalize) the amygdala response to facial emotion cues

 The authors summarize their findings “These data demonstrate that negative emotional-processing biases occur automatically, below the level of conscious awareness, in unmedicated, currently depressed people…”.  “This nonconscious processing of emotional stimuli is consistent with evidence that the amygdala contains cells that are tuned selectively to specific stimulus characteristics, facilitating early detection of biologically salient information”.

The findings from this research suggest exaggerated amygdala responses to sad faces may be a trait marker and not just due to the presence of active depression.  This might allow this trait to be studied as an endophenotype (or potential genetic marker for depression).    Normalization of this exaggerated response with selective serotoning reuptake inhibitors may provide an additional paradigm for studying the effect of new novel antidepressants.

From the clinical standpoint, this study suggests that negative emotional cues may occur below the level of consciousness in the daily lives of those with depression.  It supports clinical experience that some dysphoria may occur in response to environmental cues patients do not remember (or recognize at the time of the cue).   Further study of emotional processing is likely to advance both the research and clinical understanding of depression and other mood disorders. 


Photo of sea gull over beach at Jupiter Island, Florida courtesy of Yates Photography.

Victor TA, Furey ML, Fromm SJ, Ohman A, & Drevets WC (2010). Relationship between amygdala responses to masked faces and mood state and treatment in major depressive disorder. Archives of general psychiatry, 67 (11), 1128-38 PMID: 21041614

The Neurocircuitry of Anorexia Nervosa

Walter Kaye, M.D., Director of the Eating Disorders Treatment and Research Program at the University of California, San Diego presented a Frontiers in Neuroscience  lecture on September, 14, 2010.   The presentation was titled: “ Is anorexia nervosa an eating disorder? New insights into puzzling symptoms”.  The presentation highlighted some his recent research that has been summarized in the manuscript cited at the end of this blog post.

Dr. Kaye noted that eating disorders characterized by a phenotype of several continuous behavioral traits that appear to have some genetic contributions.  Childhood temperamental and characterological traits associated with anorexia nervosa include:

• Drive for thinness
• Perfectionism
• Harm avoidance
• Negative emotionality
• Altered interoceptive awareness
• Inhibition
• Obsessive-compulsive personality traits

Interoreceptive awareness includes a variety of body sensations such as gastric fullness, hunger and satiety.  The insular brain cortex is being recognized as a crucial brain structure in interoceptive awareness and it's integration with other brain functions. 

Some of the imaging research in anorexia nervosa (the condition’s medical coding number is 307.1) focuses on abnormalities involving appetite.  Underweight patients with anorexia nervosa appear to have altered appetite or an ability to suppress appetite to an unusually large degree.  Appetite dysregulation in anorexia nervosa includes a dislike for high-fat foods, failure to rate food as positive when hungry and a reduced aversive response to sucrose during satiation.   Several brain circuits appear to contribution to appetite and food consumption regulation including the hypothalamus, the ventral (limbic) neurocircuit and the dorsal neurocircuit.

Dr. Kaye and colleagues have performed a series of imaging studies examining appetite in those with anorexia nervosa.   He feels the anterior insula cortex might play a key role in anorexia nervosa.  This brain structure integrates with the cortex region involved in taste.  The insula also projects and interacts with key limbic structures including the amygdale, the anterior cingulated cortex and the orbitofrontal cortex.  These regions may interpret taste in a manner that integrates affective relevance, conflict monitoring and incentive learning. 

Kaye summarized that dysregulation in both the dorsal and ventral neurocircuits may contribute to anorexia nervosa.   Cortical circuits (top-down) are overengaged in anorexia nervosa resulting in high anticipatory reactivity, behavioral rigidity, and excessive worry.  This top-down overengagement is implemented through the interoceptive pathways, insular cortex amydala and anterior cingulated cortex.  The neural information  converges in the straitum where food delay receives priority over eating.

Dr. Kaye summarizes “..future imaging studies should focus on characterizing neural circuits, their functions and their relationship to anorexia nervosa.  Genetic studies might shed light on the complex interaction of molecules within these neural circuits.”

Dr. Kaye’s Eating Disorders Treatment and Research Program website is here.

Link to full pdf of the citation noted below and other UCSD publications is here.  This has several excellent figures explaining issues in neurocircuitry in anorexia nervosa.

Photo above of limbic system with cingulate cortex in pink, amygdala in blue, and hippocampus in green from 3D Brain iPad screen shot.  Courtesy of Yates Photography

Kaye WH, Fudge JL, & Paulus M (2009). New insights into symptoms and neurocircuit function of anorexia nervosa. Nature reviews. Neuroscience, 10 (8), 573-84 PMID: 19603056