SfN Atlanta: Probing Functional Deficits in Schizophrenia
29 November 2006. The neuropsychological problems associated with schizophrenia, including deficits in attention, cognition, and learning and memory, are most acutely felt as functional deficits. Auditory hallucinations, indecisiveness, delusions, and other functional problems conspire to make normal daily activities a living nightmare. Understanding the components of these functional deficits at the level of simple measurable deficits, or endophenotypes (see SRF Live Discussion), is key to the accurate, early diagnosis and management of this disease. This focus was reflected in a slide session dedicated to functional deficits at this year’s annual meeting of the Society for Neuroscience held last month in Atlanta, Georgia.
Despite having to battle with 70,000 plus raucous fans on their way to the adjacent Georgia Dome for the Atlanta Falcons, New York Giants football game, Susan Rossell from the Mental Health Research Institute, Melbourne, Australia, managed to get the session started on time. Appropriately enough, she got the ball rolling by sharing some data on auditory processing.
While there is some evidence that Heschl’s gyrus (HG) in the primary auditory cortex and the planum temporale (PT) in the secondary auditory cortex are altered in schizophrenia patients, this has been a difficult area to study because the two regions can vary tremendously in size and location among different people. Even after normalizing there can be considerable difference in residual variability in shape and location of these regions, and the more subjects studied, the poorer the overlap, said Rossell.
To overcome this limitation, Rossell and colleagues have used accurately defined regions of interest in MRI scans. By calculating the number of voxels activated in subjects presented with semantically neutral words in a passive listening task, she has been able to more accurately map these regions. Her data suggest that the HG and PT are, in fact, activated differently in schizophrenia patients, who showed a greater tendency toward symmetrical activation (in the left hemisphere, reduced activation in HG and greater activation in the PT) than normal subjects. This effect was even greater in schizophrenia patients who experience auditory hallucinations. The finding could help to pinpoint specific groups of neurons that are affected in schizophrenia or even lead to diagnostic tests for the disease.
The auditory theme was continued by Kevin Spencer, Harvard Medical School, Brockton, who addressed the problem of sound-driven EEG responses, which in chronic schizophrenia patients are typically impaired in the so-called gamma band (30 to 100 Hertz) but not at lower frequencies (see SRF hypothesis paper). The question Spencer asked is whether this “gamma driving deficit” is also present at the very onset of schizophrenia. The answer could help doctors distinguish schizophrenia from other psychiatric syndromes such as affective psychosis and bipolar disorder, which often emerge with very similar initial symptoms. Spencer and colleagues studied gamma deficits in a total of 68 volunteers, 13 first-episode schizophrenia patients, 16 first episode affective disorder patients with psychosis, and 39 healthy controls. With stimuli at 40 Hertz, both sets of patients had reduced “phase locking,” a measure of whether neurons are firing in synchrony. For the schizophrenia patients the phase locking was deficient only in the left hemisphere of the brain, whereas for affective disorder patients the problem manifested in both hemispheres. The results suggest a relatively simple way to distinguish schizophrenia from affective disorder patients in the very earliest manifestation of disease.
Patricia Pardo, University of Minneapolis, reported that a combination of neuropsychological tests and MRI measurements can differentiate schizophrenia and bipolar disorder in their earliest manifestations. She reported results from a small pilot trial to test if any combination of 22 neuropsychological variables and 23 MRI variables can distinguish the two disorders. Pardo enrolled eight healthy controls and 10 each of schizophrenia and bipolar patients to take part in the test. She used the “leaving out one subject method” to generate sets of data based on 27 subjects, then compared that data with data from the person left out and tried to assign a diagnosis in that basis. She reported that nine different sets of 12 of the variables were 96 percent correct in giving a diagnosis. Pardo claimed that by combining those sets, a fully accurate diagnosis can be achieved. But she emphasized that this study needs to be replicated. If it holds up, she said that the results argue that schizophrenia and bipolar disorder are two distinct and separable diseases, which has been somewhat of a contentious issue.
The gamma band deficiencies were also the theme of Anna Kuznetsova’s talk. Gamma band oscillations are thought to emanate from GABAergic inhibitory neurons as they fire in synchrony (see SRF related news story), and Kuznetsova, from Indiana University-Purdue University Indianapolis, has modeled how D4 dopamine receptors might modulate that activity.
There are numerous reasons to link D4 receptors with gamma band abnormalities and schizophrenia: drugs that block the D4 receptors reduce gamma activity; D4 receptors are also most abundant in the prefrontal cortex and medium spiny neurons in the striatum, two areas believed to be most affected in schizophrenia; and polymorphisms in the D4 gene have been linked to schizophrenia. But determining how the various cortical and subcortical neural networks are modulated by D4 receptors is a major challenge.
Kuznetsova suggested a new model in which activation of D4 receptors induces methylation of cell membrane phospholipids. This affects the fluidity of the membrane surrounding ion channels and increases their open/close rate, resulting in increased firing in neurons of the prefrontal cortex to gamma band frequencies.
Mihoko Otake from the University of Tokyo, Kashiwa, Japan, described another type of modeling system, one that addresses the problem of abnormal sense of agency, meaning the tendency for schizophrenia patients to attribute their actions to someone else (overattribution) or feel they have no control over their own movements (underattribution). An important part of the model is a “comparator,” a center that can compare actual sensory feedback with predicted feedback. A comparator that functions properly facilitates a normal sense of agency according to Otake’s model. Overattribution in schizophrenia may occur when the comparator is inhibited during a voluntary movement. In this case, the movement may be “overattributed” to someone else. In contrast, when a schizophrenia patient has an involuntary movement, the model predicts that the comparator has no predicted sensory feedback and therefore underattributes the movement. As these models become more sophisticated, they may be able to help scientists understand what abnormal neuronal networks are at play in schizophrenia patients with altered sense of agency.
Overattribution can even extend to the feeling that there is a totally different “presence” in the room as recently reported by Shahar Arzy and colleagues at the Swiss Institute for Technology, Lausanne (see SRF related news story). In an attempt to identify regions of the brain that might be responsible for these feelings, Arzy and colleagues have had normal, healthy subjects carry out a mental imagery task called an own body transformation. In this task patients are asked to transform themselves into the position of an image that they see in front of them, while their brain activity is measured with high-density EEG. Arzy reported that the tempo-parietal junction (TPJ) was the one region that was consistently associated with perceptual aberrations. Duration, but not strength, of TPJ activation significantly correlated with these perceptions, suggesting that in schizophrenia patients, disturbances in self-awareness may be related to prolonged activation of this region.
Neuronal networks were also discussed by Grega Repovs, Washington University, St. Louis, Missouri, who presented data on the link between working memory problems and abnormal neural connectivity. Repovs has used functional MRI to study neuronal activation during working memory tasks. She presented data on two basic approaches, studying neural activity changes over time in individual subjects, and comparing changes between subjects. A comparison of fMRI data from a set of 11 distinct cortical and subcortical brain regions from 38 normal controls and 38 schizophrenia patients suggests that in schizophrenia there is reduced and enhanced connectivity, respectively, between the right and left dorsolateral PFC and the thalamic regions. The enhanced connectivity of the left DLPFC is possibly as a compensatory mechanism, suggested Repovs. The within subject tests also indicated reduced connectivity between the DLPFC and other cortical regions during both word and facial recognition tasks.
Connectivity was also addressed by Robert Welsh, University of Michigan, Ann Arbor. Welsh used fMRI to scan the brains of 17 stable schizophrenic subjects and 15 healthy controls to compare default state activation. The default state refers to correlated neuronal activity during the resting state that decreases when attention is directed to an external task. Welsh reported that both groups of subjects exhibited the expected default state map showing activity in the ventral medial prefrontal cortex, the PCC/precuneus and bilateral parietal areas. But in addition, the schizophrenia subjects also had activation in the left amygdala/hippocampus that was significantly greater than controls and also less activation of the left lateral parietal region. The finding pinpoints areas where default state connectivity is abnormal in schizophrenic patients.
Gabriel Dichter, University of North Carolina, Chapel Hill, used the same fMRI technique to examine emotion and attention interactions in schizophrenia patients. Poor attention control and affective dysregulation are major aspects of cognitive neuroscience in schizophrenia, and researchers have begun to explore the neural circuitry that might be the root cause of problems in these domains. Dichter tackled the question using a visual oddball task, where subjects are shown neutral images, rare aversive scenes, and target images and are required to press one button when the target image is displayed and another button for all other images. The MRI images obtained from 13 control subjects replicated previous findings—that there is a dorsal/ventral difference with the ventral (including amygdala and ventral frontotemporal cortices) being activated in response to aversive stimuli and the dorsal (including frontoparietal and anterior cingulated cortex) being activated in response to standard events. The pattern was different in patients however, who had dorsal activation in response to both target and aversive stimuli, particularly the anterior cingulate gyrus, which was only activated in response to target stimuli in control subjects. The findings suggest that anomalous dorsal activations in response to aversive stimuli might mediate the effects of affective dysregulation on attentional control.
Attention was also broached by Jun Wang from the University of Georgia, Athens, who has tested schizophrenia patients for visual attention performance. Wang and colleagues used a sophisticated test where subjects viewed superimposed images of five horizontal or five vertical bars and were asked to identify width changes in a) the central bar only, b) the three central bars, and c) only the two peripheral bars. Brain activity was monitored by EEG.
Wang reported that both control and schizophrenia patients showed the same response when asked to focus attention on the three different sets of bars, namely, activation of specific occipital sensors. But in addition, schizophrenia patients also had higher low frequency (delta and theta band) brain activity and different target detection rates and reaction times. Based on the data, Wang suggested that visual attention dysfunction in schizophrenia is based in neural abnormalities that lie upstream of the sensory cortex.
Visual stimuli were also used by Henry Holcomb, Maryland Psychiatric Research Center, Baltimore, to probe spatial memory and learning in schizophrenia patients. Although schizophrenia patients are able to learn new skills much like normal healthy subjects, they tend to do so at a slower pace. To try to understand why this is, Holcomb correlated learning with brain activation as seen using functional MRI images. He tested 22 control subjects and 20 medicated schizophrenia patients in a delayed match-to-sample test using a classic A-B-A experimental design, where subjects receive no feedback in the A trials but do receive feedback in the B trials. Subjects were all given six trials, two A trials followed by two B, and then two more A.
Holcomb reported that schizophrenia patients perform just like normal controls in the A1/A2 trials and the B1/B2 trials (accuracy rates for A’s and B’s were 0.53 and 0.59, respectively, for controls, and 0.53 and 0.62 for patients). However, in the final two trials the schizophrenia subjects scored significantly lower (0.58 vs. 0.65 for controls). While in the normal subjects fMRI signals suggested that the dorsolateral prefrontal cortex and the medial temporal lobe were active in the last two trials, in the schizophrenia patients the occipital lobe was primarily active in those tests, suggesting that once feedback is removed, there is a tendency to default back to the visual system in schizophrenics.
What exactly is the cause of schizophrenia? While that is hotly debated and it is likely that there may be many different correct answers, some speakers addressed specific etiologies of the disease. Speaking to the neurodevelopmental theory, Alan Brown, New York State Psychiatric Institute, addressed the subject of prenatal exposure to infection and risk for the disease (see SRF related news story). Brown and colleagues have examined data from the Kaiser Permanente Child Health and Developmental Study cohort of more than 12,000 children born between 1959 and 1967 and for whom prenatal exposure to influenza or toxoplasmosis is documented. From the cohort, 71 cases of schizophrenia have been identified and tested for executive function. Brown reported that those who had been exposed to prenatal infection had significantly more errors in the Wisconsin Card Sort Test and took significantly longer to complete the Trails B test. The findings suggest that exposure to prenatal infection can lead to poorer executive function in later life.
Paolo Brambilla from the University of Udine, Italy, suggested that a vascular component may be added to the theory of schizophrenia. Cortical lobe atrophy has been documented in both schizophrenia and bipolar disorder, and Brambilla and colleagues have addressed whether this may be due to altered cerebral blood flow. The researchers recruited patients from the Udine and Verona region of Italy and conducted perfusion-weighted MRI scans of the brain to measure cerebral blood volume (CBV) in 54 schizophrenia patients, 15 bipolar patients, and 30 normal control subjects. They found that in the schizophrenia patients the CBV was lower in the left parietal and temporal lobes but higher in the occipital lobe than in controls, while in bipolar patients CBV was increased in the left temporal and frontal lobes and in the prefrontal/limbic system during emotional tasks. The data suggest that alterations in blood flow may contribute to the pathophysiology of both diseases.
Last but not least, Lucas Kempf, NIMH, Bethesda, Maryland, presented data on the role of genetic polymorphisms in disease etiology. Kempf used a multimodal imaging/genetic approach to address whether a single nucleotide polymorphism (SNP) that leads to a premature stop codon in the ZDHHC8 gene has any impact on schizophrenia. The gene, located on chromosome 22q11, has been linked to schizophrenia and to velocardiofacial syndrome, which often co-segregates with schizophrenia.
Kempf and colleagues tested 125 healthy volunteers in face matching and shape matching tests combined with a working memory test (the N-back test) while measuring brain activation with voxel-based morphometry. They found that volunteers who are homozygous for the SNP have significantly lower activation of the thalamus and also lower activation of the posterior cingulate which reached significance during the emotional matching test. The finding suggests that ZDHHC8 is involved in the thalamus-cingulate memory loop and warrants further investigation as a schizophrenia risk gene. ZDHHC8 is believed to be a palmitoyl transferase that may be involved in modulating synaptic strength.—Tom Fagan.