UIHC Home Page Biostatistics Information Main Page Image Processing Lab Main Page MHCRC Main Page


Functional Imaging Projects


The following projects are either ongoing or proposed. Not all projects on the list will necessarily be completed. This page is for general interest only.

PROJECT 1: Cognitive Activation Studies of Normal Controls and Schizophenic Patients
PROJECT 2: Mood Induction in Normal Individuals and Patients with Depression and Stroke Lesions
PROJECT 3: Region Specific Alterations in Cell Number and Neurotransmitter Distribution in the Doral Thalamus in Schizophrenia
PROJECT 4: Anomalous Organization and Connectivity in Autism
PROJECT 5: Use of Global Pattern Matching to Study Individual Variation in Brain Structure and Development in Normal Individuals
PROJECT 6: Rotational Invariant Anisotropic Diffusion Imaging to Evaluate White Matter Connectivity in vivo
PROJECT 7: The Effect of Hydrational Status on Magnetic Resonance Image Volume Measurements of CSF and Tissue in Polydipsic Schizophrenic Patients
REFERENCES


PROJECT 1: COGNITIVE ACTIVATION STUDIES OF NORMAL CONTROLS AND SCHIZOPHRENIC PATIENTS
Principal Investigators: Daniel S. O'Leary, Ph.D., Nancy C. Andreasen, M.D., Ph.D.
Co-Investigators: Richard Hichwa, Ph.D., G. Leonard Watkins, Ph.D., Laura Ponto, Ph.D.
Funding Status: RO1 Grant to Andreasen, Brain Imaging in Major Psychoses
Project Duration: Current RO1 funded through 1998. Renewal is pending.

Specific Aims
1. To use PET imaging with [15O]water to assess rCBF changes associated with the timing and integrative functions of the CCTCC in healthy volunteers.
2. Using PET, to identify differences in the function of the CCTCC in patients suffering from schizophrenia and healthy volunteers.

Background and Rationale
The focus of the Functional Imaging Unit, as well as the other core and research units of the MH-CRC, is on understanding the neurobiology of mental phenomena and disease processes, with a primary emphasis on schizophrenia spectrum disorders. The experimental strategy that we utilize in this pursuit is to focus first upon the study of normal individuals. Although knowledge has accumulated rapidly in recent years through PET and fMR imaging studies a great deal remains to be learned concerning the cognitive functions of the normal brain. For example, we have frequently observed that the cerebellum is active during the performance of cognitive tasks, even in situations where there is no motor activity (Andreasen et al 1997a). Understanding the role of the cerebellum in normal cognition, and the nature of interactions of cerebellar subregions with regions of the forebrain, is critical for exploring the hypothesis that schizophrenia involves cognitive dysmetria. Our PET work to date has established that patients suffering from schizophrenia have abnormal patterns of rCBF in regions of the frontal lobe, thalamus, and cerebellum during the performance of many different cognitive tasks. We hypothesize that "synchrony," or fluidly coordinating sequences of thought and action, occurs as a consequence of very rapid on-line processing and feedback between the cerebral cortex and the cerebellum, mediated through the thalamus. The substrate of synchrony is the CCTCC. In the project described below we propose to assess rCBF during tasks that will permit a more fine-grained componential analysis of the cognitive processes and neurophysiological mechanisms leading to these rCBF abnormalities. An interrelated set of PET imaging studies will assess rCBF changes associated with cognitive processes that utilize the putative timing and/or integrative functions of the CCTCC in normal volunteers and in patients with schizophrenia.

Hypotheses for Normal Volunteers:
1. Subregions of the cerebellum play a critical function in the timing of motor actions.
2. The cerebellar timing function is not limited to motor movements but is involved in any perceptual or cognitive function that requires temporal computations (e.g., time interval perception, self-initiated action or thought, use of preparatory cues, production of connected discourse).
3. The cerebellum interacts with motor regions of the frontal lobes (premotor and supplementary motor area) in tasks requiring the sequencing of motor acts.
4. The cerebellum is functionally heterogeneous with discrete subregions involved in execution, timing, and sequencing of motor acts, and with aspects of language, attention, and memory.
5. Discrete subregions of the cerebellum interact with subregions of the frontal lobe (e.g., DLPFC, orbital frontal lobe) in tasks requiring aspects of language, attention and memory.

Hypotheses for Patients:
1. Patients with schizophrenia have a primary deficit in the cerebellar timing mechanism that will be evidenced by decreased rCBF (relative to normal volunteers) in specific subregions of the cerebellum during tasks requiring temporal computations.
2. Because of this deficit, patients with schizophrenia will be impaired on any motor, perceptual or cognitive task requiring temporal computations and will show rCBF differences from normal volunteers in discrete subregions of the cerebellum during such tasks.
3. Patients with schizophrenia will be impaired on tasks requiring sequencing of motor acts and will show rCBF differences from normal volunteers in premotor and supplementary motor areas as well as the cerebellum during tasks requiring motor sequencing.
4. Patients with schizophrenia have a general deficit in cerebellar function which results in decreased rCBF in the cerebellum during a variety of cognitive tasks.
5. Cerebellar dysfunction in patients with schizophrenia results in abnormal activity in associated thalamic nuclei and subregions of the frontal lobe during language, attention and memory tasks.

Experimental Design and Methods
A total of ten studies in three conceptually-related groupings will be performed during the funding period. Because of constraints on space, and because specific details of studies may change with new findings, we describe one protocol in detail in order to illustrate our approach to the design of cognitive experiments using PET. For the remaining studies we provide a more general description of the design, as well as the hypotheses to be tested. Based on our previous experience with PET, all of the proposed experiments can be readily implemented in the PET environment with our available equipment, and all can be readily performed by both normal volunteers and patients. Each of the ten studies will be done with samples of twenty normals and twenty patients. Our experience indicates that it will be feasible to carry out this number of PET studies (approximately 80 subjects per year) in the 5 year funding period. A total of 200 healthy volunteers and 200 patients suffering from schizophrenia will be studied during the five year period. Their age range will typically be between 18 and 50, although older individuals may be studied occasionally. The normal controls will be matched to the patients for age and gender.

It is difficult, using standard PET methods, to demonstrate that co-activated regions participate in the same functional network, or to establish the specific functions of activated regions. The hierarchical subtraction strategy that is widely used in cognitive activation studies assumes that two tasks differ only in a cognitive component of interest. Although this strategy remains a viable and important tool, the validity of the technique rests upon theoretical assumptions about the cognitive processes that are involved in an activation and control task. However, recently-developed PET strategies allow assessment of specific cognitive subprocesses that are mediated by cortical, subcortical, and cerebellar brain regions. Fewer assumptions are required by the graded task approach, in which a parameter of psychological interest is systematically varied across scans and correlations are computed between the rCBF changes and the parameter manipulated (e.g., Grady et al 1994; Beauchamp et al 1997; Rao et al 1996). A refinement of the graded task technique involves examining the effects of graded changes in aspects of a task that affect one specific component of information processing. Examples include changing the frequency of stimulus presentation to examine components of a network that are sensitive to "data-driven, bottom-up" sensory processes (Rees et al 1997), changing the frequency or complexity of motor responses to determine the components of a network that are involved in response execution (Schlaug et al 1996), and changing the load in a working memory task to isolate the executive, "top-down" components of a network (Cohen et al 1997). Using these strategies we will explore three groups of motor/cognitive processes: timing and sequencing of motor activity and perception, memory and temporal computation, and attention/executive function and temporal computation.

Group A: Timing and Sequencing of Motor Activity and Perception:
This group of studies will directly assess the hypotheses that the CCTCC is involved in coordinating fluid sequences of action and thought in normal subjects, and that this system is impaired in patients with schizophrenia. Because deficits in timing are hypothesized to be a critical aspect of the cognitive dysmetria which is characteristic of schizophrenia, all of the studies in this group will contain one common condition, in which the subject must maintain a pre-defined motor rhythm by tapping with their index finger (Ivry & Keele 1986). This condition offers a relatively pure assessment of the hypothesized timing functions of the cerebellum. Having a large number of subjects in this condition (5 studies with 20 controls and 20 patients) will greatly increase the power of "seed pixel" analyses which utilize between-subject correlations to examine the covariation in blood flow between brain regions. Placing seed pixels in the cerebellar regions that play the largest role in timing in normal controls will allow the generation of correlational maps of all brain regions which participate in the timing function. Comparison of the correlational map for controls with the map for individuals with schizophrenia will allow identification of abnormal patterns of activation in the patient group.

Study 1. Production of motor rhythms and the perception of timed intervals:
This study will assess the neural mechanisms mediating externally-cued and self-paced tapping and the perception of temporal intervals. Ivry and Keele (1989) studied patients with lesions of different types (Parkinson, cerebellar, cortical and peripheral neuropathy) on two measures of timing function and found that only the cerebellar lesion group showed deficits in both the production and perception of timing. The cerebellar patients were more variable in the tapping task and less accurate in the perception task. Jueptner et al (1995) used PET to study time perception in 6 normal volunteers. They found activations in inferior cerebellum ipsilateral to the responding hand in a control minus rest subtraction reflecting finger movements. Comparison of timing and control conditions showed additional activations of cerebellar vermis and hemispheres bilaterally, reflecting a timing process. Although Ivry and Keele found that rhythm production and time perception were both impaired in patients with cerebellar damage, it is not known whether these two processes use the same cerebellar mechanisms. The relationship between cerebellar timing functions and frontal-striatal mechanisms is also currently unknown. There is some evidence that time perception in the minutes to seconds range utilizes frontal-striatal mechanisms (Hinton et al 1996) whereas time perception in the millisecond (ms) range utilizes the cerebellum. However, there has been little or no study of the interaction between cerebellar and frontal-striatal mechanisms during tasks requiring temporal computation.

The study will utilize the following eight conditions:
1. Externally-paced tapping (1 hertz [Hz]). This condition will require subjects to tap on a response key, using the index finger of their dominant hand, in pace with an external auditory stimulus at a constant rhythm of 1 tap per second. In comparison with the self-paced conditions, this condition will provide a means to "subtract out" motor mechanisms involved in response execution. It will also provide a condition on which patients with schizophrenia may perform equivalently to normal volunteers. Although some studies have found that patients with schizophrenia have abnormal rCBF during simple motor tasks (Mattay et al 1997; Schroeder et al 1994), these studies have used an unpaced finger to thumb response, and there have also been findings of normal activation patterns during simple motor tasks in schizophrenic patients (Buckley et al 1997; Weinberger and Berman 1996). 2. Self-paced tapping (1 Hz); 3. Self-paced tapping (3 Hz); 4. Self-paced tapping (6 Hz). These three self-paced tapping conditions will utilize the same basic procedure and will use the graded task approach (see above) to assess the brain regions that participate in generating motor rhythms of differing frequencies. Approximately 60 seconds prior to bolus arrival during each PET imaging session, an auditory stimulus will begin to externally-cue the subject to tap at one of the specified frequencies (i.e., 1, 3, or 6 Hz). The subject will be instructed to continue to tap at the same pace when the stimulus is discontinued, which will occur 10 seconds prior to bolus arrival, and to continue tapping until told to stop. Several functional imaging studies have assessed the effects of different rates of externally-cued finger movements on rCBF in normal volunteers (Blickenberget al 1996; Rao et al 1996; Sadato et al 1997). These studies have typically reported linear increases with faster tapping rates in sensori-motor cortex, and increased rCBF with no rate dependence in cerebellum and other motor regions. To our knowledge no study has explored the effects of different movement rates for a self-paced task or in patients with schizophrenia.
5. Time interval perception and judgment (500 ms). This condition will test perception of time in intervals (the "internal clock" that is hypothesized to be mediated by the cerebellum). We will present pairs of intervals to the subject with an interstimulus interval of about 1 s. The standard 500 ms interval will be defined by a 50 ms tone at the beginning and end of the interval, as will comparison intervals of 400 or 600 ms. Subjects will respond with a right index finger press if the second interval is perceived as shorter and with a right middle finger press if it is perceived as longer.
6. Control for condition 5. Subjects will listen to pairs of identical 500 ms intervals and alternately press with their index and middle fingers.
7. Time interval perception and production (500 ms). Immediately prior to PET imaging the subjects will be presented with multiple training stimulus intervals of 500 ms, marked by 50 ms tones at the beginning and end of the interval. During PET imaging an initial 50 ms tone will be presented and subjects will be instructed to press a button with their index finger when 500 ms has elapsed. There will be an interstimulus interval of about 1 s, and 30 trials will be collected during PET imaging. Condition 1 (cued tapping at 2 Hz) and Condition 3 (self-paced tapping at 2 Hz) will be comparison conditions.
8. Time interval perception and production (10 s). This condition will be similar to the previous one with the exception that subjects will receive training on 10 s intervals. During PET imaging they will be cued with a 50 ms tone and instructed to press a button with their index finger when 10 s has elapsed. There will be 5 trials during PET imaging with an interstimulus interval of about 1 s. This condition will be contrasted with Condition 7 to determine if the same cerebellar and frontal-striatal regions are activated by the perception of intervals in the millisecond and second range.

Hypotheses
1. The vermis and regions of lateral cerebellum will be activated during time interval perception in the ms range.
2. The production of motor rhythms will activate the same cerebellar regions as does time interval perception.
3. The frequency of motor production (i.e., tapping at 1, 2, or 3 Hz) will not change the size or amplitude of the cerebellar activation related to timing functions (vermis and regions of lateral cerebellum), but will increase rCBF in inferior cerebellum, premotor and motor cortex which are involved in motor execution.
4. The perception and production of time intervals in the multiple seconds range will cause greater activation of frontal and striatal regions and less cerebellar activation than will the perception and production of time intervals in the millisecond range.
5. Patients with schizophrenia will perform similarly to the normal volunteer group and show similar rCBF during externally-cued motor tasks, but will perform more poorly and show abnormal patterns of activation in the cerebellum during self-paced tasks and tasks requiring the perception of time intervals.

Study 2. Timing and sequencing motor movement:
This study will compare activation conditions that require simple motor movements (e.g., finger tapping) with tasks requiring progressively more complex motor sequencing. One condition will involve the self-paced tapping task (1 Hz) described above. Another condition will involve tapping the four fingers sequentially from index to small finger paced by a visual display of the numbers one through four on a video monitor. In addition to the repetitive single finger movement required in the tapping task, this task requires motor sequencing of a relatively simple nature. Another task will again be paced by the digits one through four on the video monitor, but will require the subject to perform a more complex sequence of finger movements (e.g., the index finger must be tapped twice, the middle finger once, the fourth finger three times, the little finger twice, and then this pattern is reversed). Schizophrenic patients showed abnormal activation of these areas during this finger movement task (Guenther et al 1994). The effects of learning will then be explored by having subjects practice the complex sequence during the interval between PET imaging conditions, so that rCBF can be compared during the initial performance of the complex sequence, after an intermediate amount of practice, and during the final PET imaging condition when the sequence has been well-learned. To confirm a recent report that rCBF differences during task performance are seen for many hours after practice (Shadmehr et al 1997), rCBF during the newly learned motor sequence will be compared to rCBF during a sequence that was taught to the subject and practiced extensively several days prior to PET imaging.

Hypotheses
1. Activation of cerebellar regions involved in timing functions will be progressively greater in tasks with more complex sequencing requirements.
2. Activation of frontal motor regions (premotor cortex and SMA) will be progressively greater in tasks with more complex sequencing requirements.
3. Patients with schizophrenia will show abnormal patterns of rCBF in cerebellar and frontal regions during motor sequencing tasks.
4. The pattern of rCBF in CCTCC will change during the learning of a motor sequence.
5. Patients with schizophrenia will have a more normal pattern of rCBF in CCTCC during well-practiced than during novel motor sequences.

Study 3. Stimulus-driven vs self-initiated action:
This study will examine self-initiated versus stimulus-driven actions. Frith's model of willed action proposes that frontal lobe mechanisms play a critical role in the initiation and monitoring of action, and that these mechanisms are impaired in patients with schizophrenia (Frith 1987; Frith et al 1991a). Hyder et al (1997) recently performed an fMRI study in which they replicated the dorsolateral prefrontal (DLPFC) activations due to willed action found by the earlier PET studies (Frith et al 1991b, Jahanshahi et al 1995), but found that slightly different regions of the DLPFC were activated during a self-initiated finger movement task than in a self-initiated verbal fluency task. Hyder et al argued that the higher resolution of fMRI versus PET revealed that willed action is not mediated by a unitary DLPFC neural mechanism. The study proposed here will advance knowledge concerning self-initiated action by looking at more specific components of self-initiated action, and by examining the role of the cerebellum in this process. We will compare conditions in which simple finger movements and sequences of movements are triggered by external stimuli with conditions in which subjects respond at their own initiative. We will also compare the initiation of manual motor acts with the initiation of verbal sequences. We will examine the relationship of activations in DLPFC and other cortical motor regions (sensorimotor, premotor, SMA and dorsal parietal cortices) with cerebellar activations during tasks requiring stimulus-driven and self-initiated responses.

Hypotheses
1. Regions of the lateral cerebellum as well as regions of DLPFC will show differences in rCBF during externally-cued and self-initiated responses in normal volunteers.
2. Patients with schizophrenia will show abnormal rCBF in the cerebellum as well as in cortical motor regions during the execution of motor movements.
3. The rCBF abnormalities in the patient group will be greater during self-initiated than during stimulus-driven responses and will be differentially greater in the more complex than in the simpler movement conditions.
4. Deficits will be similar in the patient group for both manual and verbal responses.

Study 4. Use of preparatory information and inhibition of motor action:
This study will assess the ability of subjects to utilize stimulus information to prepare motor responses, and will also assess subjects' ability to inhibit a motor response once it has been initiated. We will use five conditions described in a recent PET study by Deiber et al (1996), in which a preparatory stimulus provides either full, partial or no information concerning two aspects of an upcoming movement: finger to be moved (index or little finger) and type of movement (abduction or elevation). These conditions require 4 injections, and in a fifth condition the subject will make one of the four possible movements of their volition. Events will be timed to emphasize response preparation rather than execution during the critical imaging window (Hurtig et al 1994). Dieber et al found some brain regions that were activated during all of the preparatory conditions, suggesting a single anatomic substrate for motor preparation independent of information context. Other brain regions were differentially activated during preparation for self-initiated movement and during the full and partial information conditions, giving evidence of ties between aspects of movement preparation and specific brain regions. Inhibition of motor movements will be assessed using a variant of a task described by Dejong et al (1996). A warning signal is presented followed by a stimulus that requires a response. We will use an arrow pointing to the left or right as a stimulus to indicate a left or right move of the joystick. In the Dejong et al study, an auditory stop signal (1000 Hz, 50 ms tone) is presented on 40% of the trials at differing intervals following the response arrow. We will maintain this percentage of stop signals over the course of a 2 minute task, but will increase the density of stop signals to 80% during the critical imaging window.

Hypotheses
1. We will replicate the brain regions found by Dieber et al to be activated during response preparation in normal volunteers.
2. Patients with schizophrenia will have abnormal rCBF patterns during all of the response preparation conditions indicating a deficit in the anatomic network mediating preparation.
3. Schizophrenics will also show abnormal rCBF patterns during the condition requiring preparation for volitional movements.
4. The cerebellum and other components of the CCTCC will be involved in activation during movement inhibition in normal volunteers, and abnormal activation will be seen in the patient group.

Study 5. Conscious and unconscious perception:
Although consciousness may be difficult to quantify, patients with schizophrenia often note that abnormal conscious experiences are one of the most salient and frightening aspects of their disorder. The study proposed here represents an initial attempt to explore brain mechanisms underlying differences in conscious experience in normals and in patients with schizophrenia. In order to explore conscious phenomenon we will exploit the fact that visual stimuli that are presented at too fast a rate to be consciously perceived may nevertheless show evidence of being processed at a meaningful level. We will compare conditions in which meaningful and non-meaningful stimuli are presented at rates fast enough to preclude conscious perception. Two classes of stimuli will be used -- familiar and unfamiliar faces, and real and nonsense objects. For the proposed PET study we will use a titration procedure in a preliminary session to determine a visual presentation speed for each individual that is too fast for conscious perception of objects and faces. Comparison conditions will involve presentation of similar stimuli at rates that are clearly perceptible. Although the role of the cerebellum in perception is less well explored than is its role in motor function, our own studies as well as those of other groups have shown that the cerebellum is activated during the processing of faces as well as in other tasks requiring visual perception (Andreasen et al 1996f; Fiez et al 1996). Given the role played by the cerebellum in timing functions, it seems likely that the cerebellum will be more evenly activated by stimuli that are presented very fast than by stimuli that are presented at slower rates.

Hypotheses
1. Different occipital and temporal lobe regions will be activated by the presentation of lists of real objects and lists of familiar faces at rates that are clearly perceptible (e.g. one stimulus per second).
2. Presentation of familiar face and object stimuli at rates that are too fast for conscious perception will result in similar patterns of activation in occipital and temporal lobe regions (i.e., there will be different patterns of activation for faces and objects but these patterns will be similar to those seen at slower rates of presentation of each stimulus type).
3. The cerebellum will be more activated by fast than by slower rates of stimulus presentation.
4. Although the stimuli are not consciously perceived, unfamiliar faces presented at a very fast rate will show a different pattern of activation than will familiar faces presented at the same rate. Difference will be seen as well between activation patterns evoked by real and nonsense objects presented at rates too fast for conscious perception.
5. Individuals with schizophrenia will show abnormal patterns of activation for stimuli presented at rates that preclude conscious perception.

Group B: Memory and Temporal Computation:
This group of studies will continue our studies of memory and assess the hypotheses that the timing functions of the CCTCC are involved in aspects of memory performance, and that patients with schizophrenia will show greater impairment on memory tasks that rely more heavily on these timing functions.

Study 6. Recognition vs recall memory:
This study will compare recognition and recall memory for materials studied one week, one day or immediately prior to PET imaging. This will allow us to assess changes in activation that may be due to consolidation of the material into long-term memory (or alternatively, to forgetting). We will use verbal materials that have been shown in previous cognitive studies (Tulving 1983) to be highly memorable after only one or two exposures (e.g., "an exercise for the jumpy--trampoline"). Subjects will be exposed to unique sets of 30 such phrases at one week, one day, or approximately one minute prior to imaging. In order to prevent rehearsal of the items presented to the subjects one week and one day prior to imaging, the subjects will be asked to rate each item for "meaningfulness" rather than to try and memorize it. For each list of 30, 15 items will be tested via recognition (the stem i.e., "an exercise for the jumpy" will be on the video screen at the top, and the correct item and a foil on the screen below--subject moves a finger switch toward the correct item), or recall (the stem is again at the top of the screen, "yes" and "no" are beneath--the subject moves finger towards "yes" (I remember the answer) or "no" (I don't remember). For recall conditions subjects will be asked to say the item after imaging is over to check the honesty of their manual responses. A sensorimotor baseline condition will have a sentence stem with two words beneath it - a foil and a word that is in the sentence stem. Subjects will be instructed to move their finger towards the repeated word. This baseline task has the same reading requirements and response as the recall and recognition conditions but no memory is needed. A second baseline condition will have a sentence stem and an answer that the subject can retrieve from memory without having been pre-exposed (e.g., a pointy boat paddled by Indians). This will be a recall condition--the subject responds manually "yes/no" during PET and is asked for the correct word after imaging.

Hypotheses
1. Recall tasks will activate the CCTCC to a greater extent than will recognition tasks at all three exposure intervals.
2. Materials presented at one week and one day prior to imaging will activate the CCTCC to a greater extent than will the materials presented immediately prior to PET imaging.
3. Patients with schizophrenia will show abnormal rCBF in all memory conditions, but will be more normal during recognition than during recall tasks and more normal for the immediate memory conditions than for the longer-term conditions.

Study 7. Strategy, effortful search, and retrieval success (ecphory):
Moscovitch (1989) has distinguished two types of memory search. One, associative or cue-dependent, depends on a neuronal network that includes the hippocampus. A second, strategic process requires the activation of a circuit that includes the frontal lobes and presumably other components of the CCTCC. The Toronto PET group (Tulving et al 1995, Kapur et al 1994a, Nyberg et al 1995) has proposed that the right frontal lobe is differentially activated by any task that requires retrieval (i.e., frontal activation results from "retrieval mode"), whereas posterior brain regions are activated during successful retrieval. The Hammersmith group has argued that the frontal lobes are more involved in successful retrieval than in the effort of retrieval (Rugg et al 1996). Our own studies have provided support for the importance of frontal lobe for strategic retrieval (Paradiso et al in press), but are also consistent with a role of the right hemisphere in retrieval and the concept that successful retrieval activates posterior brain regions (Andreasen et al 1995e, f). The study proposed here will attempt to disentangle the effects of effort, strategy and retrieval success in memory. The study will have four conditions with differing levels of success and effort i.e., high effort with high success, high effort with low success, low effort with high success, and low effort with low success. A second set of conditions will compare a task that is aided by a self-generated memory strategy (e.g., remembering word lists containing categories, such as tools, food, or clothing, is greatly aided by clustering related items) with a task in which strategies are less useful (e.g., remembering word lists that have no internal structure). We will compare a condition in which subjects are told what strategy to use with the two previously described conditions.

Hypotheses
1. Any task requiring "retrieval mode" will activate the right DLPFC, and successful retrieval will activate posterior brain regions (precuneus and cerebellum).
2. Greater effort will cause a similar pattern of frontal activation as does strategic retrieval i.e., both effortful and strategic retrieval will bilaterally activate the orbital frontal lobe.
3. Greater retrieval success will cause greater activation of the right DLPFC and posterior brain regions than will less retrieval success.
4. Individuals with schizophrenia will show less orbital frontal activation concurrent with less spontaneous use of strategies.
5. Even when given a strategy the schizophrenic group will show less activation of orbital frontal lobe and CCTCC than will the control group.

Study 8. Declarative versus implicit memory: This study will use PET to evaluate the interaction between cerebellar brain regions which are thought to be essential for perceptual-motor learning and memory, and hippocampal and neocortical systems that are thought to mediate declarative memory. PET imaging will be used to test specific predictions concerning the interactions between regions of the cerebellum and other brain regions during the acquisition and extinction of the classically-conditioned eyeblink response in normal volunteers and in patients with schizophrenia. The PET conditions will be as follows: 1) Fixation baseline task; 2) Non-paired, presentation of tones; 3) Classical conditioning, early acquisition phase; 4) Classical conditioning, mid-acquisition phase; 5) Classical conditioning, late acquisition phase; 6) Early extinction phase; 7) Mid-extinction phase; 8) Late extinction phase.

Hypotheses
1. Because long-term depression during learning has been seen in animal studies (Ito 1984), rCBF should decrease in cerebellar cortex during learning.
2. In contrast, rCBF should increase in deep cerebellar nuclei during learning.
3. Regional CBF will increase in the hippocampus during initial learning, but decrease to baseline levels as asymptotic learning levels are achieved.
4. Regional CBF will also change in the anterior cingulate and basal ganglia during conditioning, but available data do not permit predictions of the direction of change.

Group C: Attention/Executive Function and Temporal Computation:
This group of studies will assess the role of the CCTCC in higher-order cognitive functions that rely upon top-down mechanisms traditionally thought to be mediated by frontal and parietal cortex with little or no participation of the cerebellum and thalamus.

Study 9. Spatial working memory and spatial attention:
The performance of tasks requiring spatial working memory has been shown to be dependent in monkeys and humans upon prefrontal neurons. Goldman-Rakic (1994) has recently suggested that working memory deficits may be a core neuronal deficit underlying the symptomotology and cognitive deficits of schizophrenia. Spatially-directed attention has also been studied in both primates and humans and abnormalities in bottom-up attentional processing have also been proposed to explain the cognitive and phenomenological abnormalities of schizophrenia. There have been recent suggestions that spatial working memory and spatial attention may interact, and/or be part of the same cognitive-neural system (Desimone et al 1995). There have also been suggestions that deficits in spatial working memory and spatial attention may be a core problem in schizophrenia. This study will involve a direct comparison of spatial working memory and spatially-directed attention. The following conditions will be counterbalanced across subjects: 1) Spatial working memory. 2) Spatial attention. 3) Sensorimotor control task. 4) 2-back spatial task. 5) Shifting attention task. 6) S-M control task with multiple button presses. 7) Condition #1 but with background flashing at twice the rate as in #1. 8) Condition #2 but with background flashing at twice the rate as in #2.
Hypotheses
1. Spatial working memory and spatial attention will activate different anatomic circuits although some brain regions will be activated by both tasks. Spatial working memory will activate DLPFC, superior parietal cortex and cerebellum, whereas sustained spatial attention will activate overlapping regions of DLPFC and superior parietal lobe and the pulvinar nucleus of the thalamus, but not the cerebellum.
2. Changing the rate of stimulus presentation will affect occipital and temporal lobe activations for the spatial attention but not the working memory task.
3. The shifting attention task and the 2-back task will activate similar regions of the cerebellum, but other components of the CCTCC will differ between the tasks.
4. Patients with schizophrenia will show less activation in CCTCC and additional activations not seen in controls in both the spatial attention and spatial working memory tasks.

Study 10. Executive function and motor execution:
The working memory paradigm is now one of the most widely used tasks in functional imaging and has consistently been shown to activate the DLPFC (e.g., Smith et al 1996, McCarthy et al 1996). The most widely cited model memory (Baddeley 1986, Shallice 1988) argues that a modality-free "central executive" utilizes modality-specific "slave"-systems to maintain information. The executive has been proposed to be localized in prefrontal cortex. The wealth of information available at multiple levels of analysis concerning working memory allows the working memory paradigm to be used as a tool for the functional parcellation of information processing tasks and their neural substrates. The study proposed here will use a graded task approach to independently manipulate the memory load on the executive (e.g., remember 3, 4 or 5 items in a Sternberg paradigm) and the complexity of the motor response that is required (e.g., single finger tap response versus 2 finger sequence versus 4 finger sequence). This approach will differentiate the regions of frontal cortex and cerebellum that are involved in executive memory functions from frontal and cerebellar regions that are involved in aspects of response execution.

Hypotheses
1. Regions of DLPFC and cerebellum that are activated during working memory will show linear increases in rCBF with increasing memory load in normal volunteers. Increases in memory load will not affect frontal and cerebellar regions involved in response execution.
2. Different regions of frontal lobe and cerebellum will be activated by motor response execution and these regions will show linear increases in rCBF with increasing response complexity in normal volunteers. These regions will not be sensitive to differences in memory load.
3. Patients with schizophrenia will show abnormal patterns of activation in brain regions involved in executive function and response execution and the abnormalities will be greater at higher levels of difficulty.

MR Acquisition:
As described above. All PET images are registered on MR images using the AIR method (Woods et al 1993; West et al 1997; Studholme et al 1997).

Data Analysis:
Papers in the Appendices discuss many of the relevant issues in data analysis of PET data, including sample size and choice of statistical methods. In general, we plan to continue to use exploratory FOI analyses in the normals as the initial step in data analysis, applying our adaptation of the Montreal method (Worsley et al 1992; Arndt et al 1995a). For between group comparisons we will use our locally-developed randomization method (Arndt et al 1996c). Correlational analyses using our seed pixel method with external vectors will be used for the experiments involving graded stimuli (Clark et al 1984; Worsley et al 1995; Strother et al 1995; Friston et al 1993).

Back to top

PROJECT 2: MOOD INDUCTION IN NORMAL INDIVIDUALS AND PATIENTS WITH DEPRESSION AND STROKE LESIONS
Principal Investigator: Robert G. Robinson, M.D.
Co-Investigators: Sergio Paradiso, M.D., Richard Hichwa, Ph.D., G. Leonard Watkins, Ph.D., Laura Boles Ponto, Ph.D.
Funding Status: RO1 Grant to Robinson
Project Status and Duration: Funded 03/31/96-04/01/01

Specific Aims
The overall aim of this proposal is to investigate whether specific neural structures are involved in the physiological and psychological processes leading to emotion. Specifically, we will:

1. Investigate physiologic patterns of brain response (EEG) to emotional activation in normal individuals and in patients with single vascular lesions involving orbital frontal or basal temporal cortex or basal ganglia or dorsal medial thalamus, (the ventral-lateral limbic circuit) and compare them with patients having lesions of parietal or posterior temporal cortex or brainstem, (non-limbic structures) to determine whether the lesions interfere with either the psychological or physiological response to an emotional stimulus i.e., happiness, sadness, or fear.

2. Examine the pattern of physiologic brain response (EEG and PET) in a subset of the above subjects. The subjects would be stroke patients with thalamic (limbic) lesions or brainstem (non-limbic) lesions and matched normals.

Measures of autonomic nervous system (ANS) function (pulse, heart rate, skin conductance, finger temperature, finger pulse amplitude, and pulse transmission time to finger and ear) will also be evaluated in patients with or without limbic lesions.

Background and Significance
This project investigates the neuroanatomic and neurophysiological substrates of emotion through PET imaging and electrophysiological studies of normal individuals and individuals with discrete vascular lesions involving orbital frontal or basal temporal cortex or basal ganglia or dorsal medial thalamus, (the ventral-lateral limbic circuit). Patients having lesions of parietal or posterior temporal cortex or brainstem, (non-limbic structures) will also be studied to determine whether the lesions interfere with either the psychological or physiological response to an emotional stimulus i.e., happiness, sadness, or fear. This investigation of the effects of localized limbic and non-limbic lesions on rCBF during the processing of emotional stimuli will provide extremely valuable comparison data for a study we have recently completed of rCBF and anhedonia in schizophrenic patients. In this study, individuals with schizophrenia and normal controls were exposed to pleasant, unpleasant, and neutral visual and olfactory stimuli to explore one of the fundamental negative symptoms of schizophrenia.

Hypotheses
1. Following lesions of the orbitofrontal cortex, polar basotemporal cortex, striatum or dorsal medial thalamus (ventral-lateral limbic circuit), patients will not experience fear or sadness (with right lesion), or happiness (with left lesion), or show the EEG correlates of these emotions, but they will maintain their physiological response (ANS) unless the amygdala input to the hypothalamus is destroyed.
2. Lesions of posterior temporal cortex, parietal cortex and brain stem (non limbic lesions) will not interfere with either the psychological perception or the physiological response to positive or negative emotions, except when the lesion impairs input to the limbic cortex (i.e., some patients will have a preserved emotional perception, but a faulty activation such as in pathological laughing and crying).
3. Normals will activate one or more portions of their left limbic system in response to happy emotional processing and the right limbic system with sadness and fear. Sadness will involve a frontal flow increase and fear a temporal flow increase.
4. That patients with lesions of the thalamus (limbic circuit) will not experience emotion processed by the side of the lesion and will not, therefore, increase orbitofrontal or basotemporal-amygdala blood flow in response to the presented emotional stimuli, however, patients with brainstem lesions (also posterior circulation lesions) will experience emotion and activate limbic areas.

Preliminary Studies
In a preliminary study, a sequence of films was shown with another neutral (baseline) film, followed by happy, fearful, neutral, fearful, happy, neutral film clips. Immediately after each activation condition, the patient reported the intensity of emotional response using an analogue scale. Facial expression was recorded via videotape during activation/scanning utilizing a mirror placed inside the scanning ring. We found a significant effect of emotional activation upon regional blood flow. Using image subtraction analysis (i.e., emotional activation state minus the combined neutral conditions), there were significant findings which were reproducible (there were two activations of the same emotion in each subject). There was an area in the left prefrontal region that was activated during the viewing of happy images that was not activated during fearful images (50% increase), and an area of the right superior temporal gyrus that was activated in fear/disgust that was not activated during happy images (70% increase).

Methods
PET imaging data, EEG data and ANS data will be collected from subjects while they are undergoing an activating procedure designed to stimulate emotion. Emotional activation will involve three types of emotion: happiness, sadness, and fear. Emotions will be elicited using short (average 60 sec.) segments from motion pictures. They are presented to the subjects as video clips without sound. They have been selected on the basis of their ability to elicit EEG evidence of emotional activation, and have been validated in previous studies. Psychological perception (i.e., subjective report of the patient) will be recorded at the end of each video presentation to determine emotional responses.

Patients who have had strokes with single lesions which are identifiable on MRI or CT scan will be selected based on the criteria in Specific Aims 1 and 2 above. Since lesions are only rarely restricted to small specific structures, we will select from over 400 patients per year to find those with relatively discrete lesions (e.g., limited primarily to one specific region such as basotemporal cortex or thalamus). The selection will involve a neuroradiologist blind to any clinical data identifying a lesion as involving a structure under study (e.g., thalamus or basal ganglia). Extension to non-study structures will be allowed if the majority of the lesion is in the desired structure (e.g., orbital frontal lesion with extension to dorsal lateral cortex or white matter) as long as two study structures (e.g., frontal and temporal cortex) are not involved. The subjects to be used in these studies will also be only non-depressed, non-anxious patients (as determined by not fulfilling symptom criteria for major or minor depressive disorder or generalized anxiety disorder) and without history of having met these diagnostic criteria in the past. They will be studied approximately 3 months post-stroke when the acute medical complications and lesion characteristics have stabilized.

There will be 120 patients with stroke studied and 20 normal control subjects with a subset of 60 patients undergoing PET imaging. Numbers of patients in each category will be as follows (total 140 recruited over 5 years).

Specific Aim #1 (EEG and ANS only): Inferior frontal cortex, 20 (10 left and 10 right), Anterior temporal cortex 20 (10 left and 10 right), Thalamus 20 (10 left and 10 right), Basal ganglia 20 (10 left and 10 right), Brainstem 20, Parietal-posterior temporal 20 (10 left and 10 right), Normal Controls 20.

Specific Aim #2 (PET Study): Of the subjects above, 60 subjects will have a PET study in addition to their EEG and other studies. Thalamus 20 (10 left and 10 right), Brainstem 20, Normal Controls 20.

PET Imaging: PET imaging will utilize the methods noted above. The sequence of conditions will be neutral (baseline), happy, fearful, neutral, fearful, happy, neutral.

Autonomic Nervous System Response: During and following emotional activation, measures of blood pressure, pulse, galvanic skin conductance and facial expression will be recorded. They will be collected in a standardized way using procedures demonstrated in previous work to produce measurements which are reliable and valid. Facial expression will be recorded on videotape.

EEG: Quantitative EEG measurements will be taken during the emotion activating procedure. EEG measurement data obtained from the subjects will be analyzed by Dr. Davidson in Wisconsin. He will be blind to other experimental data. To permit computation of reference-free montages, EEG will be recorded from 30 scalp locations. In addition, two channels of EOG will be recorded (to obtain measures of vertical and horizontal eye movements) and a trigger channel will denote the onset and offset of data acquisition periods.

Other Measures:
All patients will undergo a standardized neurological examination using the National Institute of Health (NIH) Stroke Scale. Patients will be examined by a psychiatrist using a semi-structured interview, the Present State Exam (PSE) including a modification to assess lifetime diagnosis. Quantitative evaluation of social functioning will be done using the Social Functioning Exam and the Social Ties Checklist. The Johns Hopkins Functioning Inventory will be used to assess the patient's ability to perform activities of daily living. Assessment of cognitive functions will be done using a neuropsychological test battery designed in collaboration with Dr. Edith Kaplan at the Boston Veterans Administration Hospital.

Data Analysis:
We recognize that this study will generate a huge amount of data, and that data analysis will represent one of the most formidable aspects of this study. We have, therefore, put together a team that has a track record in this kind of data analysis. Dr. Stephan Arndt, Director of the Biostatistical Core Unit of the MH-CRC will serve as consultant for image data analysis. Dr. Richard Davidson at the University of Wisconsin at Madison will provide instrumentation for recording EEG and ANS data, as well as data processing and analysis at his lab in Wisconsin. Dr. Davidson is highly experienced in data reduction of EEG studies. He will work closely with Dr. Arndt to implement EEG-PET co-registration analysis. Data analysis will examine the relationship between subjective experience of emotion, (from the post video ratings) behavioral (i.e., facial expression) and biological (i.e., autonomic measures and EEG measures) concomitants. This will be done both cross-sectionally (e.g., at the end of the film does the patient's report of sadness and EEG frontal asymmetry correlate?) and longitudinally (e.g., is there a frontal asymmetry prior to a change of mood rating?). The analyses will examine subjective experience of emotion (e.g., sadness) as the dependent variable and autonomic and electrophysiological measures as the independent variables. Logistic (for categorical) and linear (for continuous measures) relationships of emotional and biological variables as predictors of one another and temporal concomitants will be carried out. In addition, we will examine if there is an effect of the presentation order of the activating video clips. Groups will be formed on the basis of lesion location in areas already described. We will control for differences between groups in lesion volume and cortical versus subcortical effects (i.e., analysis of variance will test for independent and interaction effects). We will examine effects of age, sex, education, cognitive impairment and the other variables as previously described. We anticipate that most of these variables will be controlled across groups and that multiple factors will not have to be controlled. This has been our experience in prior research in this area.
Back to the top

PROJECT 3: A POSITRON EMISSION TOMOGRAPHY STUDY OF LANGUAGE PRODUCTION IN ADULT AUTISTIC INDIVIDUALS
Principal Investigator: Joseph Piven, M.D.
Co-Investigators: Daniel S. O'Leary, Ph.D., Richard Hichwa, Ph.D.
Funding Status: This project is currently not funded
Project Status and Duration: Ongoing. ROI submission planned 1998.

Specific Aims
This project employs PET imaging with [15O] water to investigate rCBF during the performance of cognitive tasks in adults who were diagnosed in childhood as autistic (hereafter called adult autistics). Adult autistics will be imaged with PET during the performance of cognitive tasks which are typically performed abnormally by these subjects, and their rCBF will be compared both to normal volunteers and to patients with schizophrenia who have performed the same tasks.

Background and Rationale
Schizophrenia and autism share many features in common, which include symptoms such as abnormal social interactions, and cognitive abnormalities such as atypical language use. Both schizophrenia and autism are disorders that are clearly linked to brain abnormalities, although the exact nature of the brain dysfunction underlying either disorder is currently unknown. Autism has a clear developmental course which suggests abnormalities in brain maturational processes. Neuropathological studies have demonstrated abnormalities in the cerebellum, and medial temporal lobe structures. A recent MRI study performed here at the University of Iowa revealed that adult male autistics had larger brain volumes than did a group of age-matched control males (Piven et al 1997a). The increase was observed both in brain tissue volume and in the ventricular system. Study of brain function with PET and [15O]water during the performance of specific cognitive tasks may help to determine the functional consequences of the developmental abnormality that characterizes autism. Specifically, the present study will allow determination of whether brain regions activated by cognitive tasks in autistics are the same or different than areas activated in normals and in patients with schizophrenia. There has been relatively little study of brain function in autistic individuals. A recent review of PET studies performed with autistic subjects (Horwitz & Rumsey 1994) notes that no focal abnormalities in metabolism have been found. However, PET studies performed to date have assessed glucose metabolism in either a resting state or during low-level cognitive tasks utilizing [18F]fluorodeoxyglucose (FDG). Horwitz and Rumsey (1994) suggest that the use of more informative activation tasks during PET imaging is likely to be a productive strategy for assessing brain function in autism. PET with [15O]water permits the use of repeated back-to-back studies with differing activation tasks, allowing the possibility of isolating specific cognitive impairments.

Hypotheses
1. Adult autistics will demonstrate rCBF abnormalities (i.e., differences from the pattern exhibited by the normal group) during the performance of language production tasks.
2. The rCBF abnormalities of the adult autistic group will have both similarities and differences to those exhibited by the schizophrenic patient group.
3. The adult autistic group will show a unique pattern of rCBF abnormalities on the higher-level language production tasks requiring production of an imaginative story concerning the cognitive state of others (i.e., rCBF correlates of deficits in imaginative narrative and "theory of mind" tasks).


Preliminary Studies
We have performed PET imaging studies on a total of 3 adult autistic subjects. Data analysis for this study requires averaging across at least 5 subjects, and we plan an initial, preliminary analysis after 2 more subjects have been imaged. This is the first time we have attempted PET imaging with adult autistics and in practical terms we have learned that the subjects tolerate the procedure well, although they require very concrete information about each step in the procedure. The cognitive activation protocol that we used is one with which we have already collected data on a sample of 13 controls and 25 patients. The data from the schizophrenia patient group has not yet been analyzed. Analysis of the rCBF changes in the normal volunteer group that were associated with the activation tasks below yielded the following findings: 1) Reading single words caused increased rCBF (over a baseline task) in visual cortices in the occipital lobe, in auditory cortices in the left and right superior temporal gyrus, and in the left frontal motor strip. 2) The tasks in which normal subjects produced verbal narratives all showed increases in rCBF in the cerebellum, left superior temporal gyrus, left thalamus and regions of the left frontal lobe. Differences between the narrative conditions were largely within the left hemisphere, in the insula, inferior frontal lobe and inferior temporal lobe.

Methods:
Patient Sample: Twenty-two patients will be selected from a larger sample who have been assessed by Dr. Piven. The patients selected will be males above the age of 18 years, who have IQs above borderline. The patients selected have already had a usable MR scan, using the same T1 weighted sequence used by the schizophrenic research project.

The procedures to be used for PET image acquisition are as described above. The tasks are designed to require differing types of linguistic processing, beginning with simple reading, and progressing in stepwise stages to relatively complex tasks requiring the creation of narratives. The activation tasks were designed in collaboration with Dr. Andreasen's research team. The specific conditions are: #1 Read words. #2 Read words in sentences. #3 Read words in narrative. #4 Eyes-closed rest. #5 Non-imaginative neutral narrative, i.e., "Tell what you did after you woke up yesterday morning.") #6 Imaginative neutral narrative, (i.e., "Imagine that you are shopping in a grocery store and see Mickey Mouse pushing a grocery cart. Tell a story about what you did.") #7 Imaginative empathic story, (i.e., "Imagine that you meet a stranger who is crying. Tell a story about what happened to her.") #8 Generate a story using 5 specific items.

Data Analysis:
Within the autistic group the analysis will include both a function of interest, t-statistic mapping approach, and a region of interest-bases approach assessing rCBF within volumes traced on co-registered MRIs. Randomization analyses (Arndt et al 1996c) will be used for between-group analyses.

Back to top

PROJECT 4: CHRONIC MARIJUANA USE, BRAIN AND COGNITION
Principal Investigator: Robert I. Block, Ph.D.
Co-Investigators: Daniel S. O'Leary, Ph.D., Richard Hichwa, Ph.D.
Funding Status: RO1 grant to Block
Proposed Duration: Funded 4/97-8/99

Specific Aims
The long-term objectives of this project are to understand effects of chronic use of marijuana on brain structure, and function and cognition. The specific aim is to compare rCBF and cognition in chronic marijuana users who have limited experience with use of other drugs to non-using controls.

Background and Significance
Many studies have examined effects of chronic marijuana use on human cognition (Pope et al 1995) but only one (Block and Ghoneim 1993) has matched marijuana users and non-users on their intellectual abilities prior to the onset of drug use, to control for possible premorbid differences in cognitive abilities. Four studies examined chronic marijuana users by air encephalography (Campbell et al 1971) or computed tomography (Co et al 1977, Hannerz and Hindmarsh 1983, Kuehnle et al 1977). Measurements of the lateral and third ventricles provided evidence of cerebral atrophy in the 1st study. The other 3 studies observed no abnormalities. Several studies by Matthews and colleagues have examined rCBF of chronic marijuana users using the 133Xenon inhalation technique. One study observed no differences between chronic marijuana users and controls in rCBF or overall CBF (Mathew et al 1986a; Mathew and Wilson, 1992). A second study found that experienced marijuana users showed the lowest overall CBF values. Other researchers observed overall CBF decreases, in the absence of rCBF differences, in chronic marijuana users relative to controls (Tunving et al 1986). The discrepant results may have been due to differences in subjects' characteristics: Ss in Mathew et al's studies were volunteers without obvious problems associated with marijuana use, whereas the other researchers studied patients who had recently entered treatment for marijuana use. Subcortical changes were found in the one PET study on this topic, which used 18F-2-fluoro-2-deoxyglucose as a tracer to measure regional cerebral glucose utilization of chronic marijuana users and controls (Volkow et al 1996c). At baseline, chronic users showed decreased relative metabolism compared to controls in the cerebellum, but nowhere else. Administration of 2 mg THC increased cerebellar metabolism in both chronic users and controls. Chronic users also showed increased metabolism in prefrontal cortex, orbitofrontal cortex, and basal ganglia, whereas controls did not. The authors suggested that increased metabolism in the latter two areas could be related to the drive to self-administer the drug. This important study indicated that the brain area most sensitive to chronic effects (the cerebellum) was also the most sensitive to acute effects. Subjects in all of these studies were evaluated during "rest," i.e., undirected mental activity.

Hypotheses:
Compared to matched non-users, chronic marijuana users will have:
1. reduced or otherwise altered patterns of rCBF during performance of cognitive activation tests assessing: 1) memory retrieval for well-learned and novel materials; and 2) selective attention;
2. reduced or otherwise altered patterns of rCBF during undirected mental activity and performance of control tests (auditory reaction time and counting);
3. mild cerebral atrophy or other changes in brain structure, assessed by MR imaging and analyzed with innovative image averaging techniques;
4. selective deficits in memory retrieval and impairments of verbal expression and quantitative abilities, as we found in previous research; and deficits in selective attention.

Progress Report:
We examined effects of chronic use by comparing marijuana users and non-users who were matched on intellectual functioning before the onset of marijuana use, i.e., on their Iowa Test (ITBS) scores during the 4th grade (Block & Ghoneim 1993). Under drug-free conditions, the 12th grade versions of the ITBSs were administered along with computerized cognitive tests to determine if chronic marijuana users were impaired relative to non-users and, if so, whether impairments depended on the frequency of marijuana use. Two of the 12th grade ITBS showed impairments in heavy marijuana users relative to non-users. On a verbal memory test heavy users showed impairment relative to non-users in long-term retrieval for high but not low imagery words. Heavy users scored lower percentages of correct responses than intermediate users or light users for more difficult concepts in a concept attainment task.

Methods
Sample: Subjects will be 30 heavy marijuana users and 30 non-using controls recruited primarily by newspaper advertisements. The groups will be matched on mean scores on the 4th grade ITBS and as well as possible on education, age, height, gender, race, recent chronic use of tobacco, and their own and their parents' socioeconomic status. Design: Following an initial screening session subjects will be scheduled for 2 overnight inpatient hospitalizations in the University of Iowa Clinical Research Center, separated by a 1-week interval. A biographical interview, MR imaging, and a cognitive (non-imaging) test session will be conducted during the 1st hospitalization. The PET session with cognitive activation tests, preceded 24 hours earlier by practice on these tests and relearning of a word list used in one of the cognitive activation tests, will be conducted during the 2nd hospitalization. To verify abstention from marijuana and other drugs, blood and urine samples will be obtained at screening, on the days of admissions to the Clinical Research Center, and before the cognitive test session and PET session. Saliva samples will be obtained at admission and every 4 hour thereafter during both hospitalizations, and breath testing for alcohol with an Alco-Sensor III unit will be done at the same times.

Cognitive Activation Tests. Cognitive activation tests during the image acquisition periods will be as follows: Free Recall: In 3 memory tests, Ss will be instructed to recall as many words from a list as they can, in any order. 1) subjects will be asked to recall the previously learned word list. 2) subjects will again be asked to recall list 1, but, as in our previous PET study (Andreasen et al 1995f), the word list will be presented to subjects auditorily at a rate of 1 word per sec immediately before recall begins. Recall of a novel word list will be tested. Active Baseline for Free Recall: The active baseline will consist of repeatedly counting from 1 to 3 at a rate of approximately 1 number per sec. Dichotic Listening Test: The attend-right and attend-left dichotic listening task with CVCs, and the baseline task described in O'Leary et al (1996b, c) will be used. Rest: subjects will be given no specific instructions about mental activities, being asked simply to lie quietly with eyes closed.

Data Analysis
The quantitative PET blood flow images and MR images will be transferred to the Image Processing Laboratory of the University of Iowa Mental Health Clinical Research Center for processing. The MR processing will produce "average brains" of marijuana users and controls, which can represent their differences as an effect size map (Andreasen et al 1994k). To check for individual abnormalities that might be missed by the automated image averaging process, the MR images will also be interpreted clinically by a neuroradiologist (Dr. Yuh). FOI analysis of PET images will utilize our standard image subtraction approach. Randomization analyses (Arndt et al 1996c) will be used for between-group comparisons of chronic users and matched controls. Manual tracings of ROIs will be made on each subject's co-registered MR images. ROIs consist of left and right prefrontal, posterior inferior frontal, insula, anterior and posterior cingulate; and left and right anterior inferior temporal, thalamus, and cerebellum. The hippocampus will be an additional ROI for the memory tests. The blood flow for each ROI, expressed in ml/min/100 g, will be submitted to ANOVAs including group (marijuana users vs. controls) and condition (the pairs of conditions used for subtractions in the corresponding FOI analyses).
Back to top

PROJECT 5: ACUTE MARIJUANA EFFECTS ON REGIONAL CEREBRAL BLOOD FLOW
Principal Investigator: Daniel S. O'Leary, Ph.D.
Co-Investigators: Robert I. Block, Ph.D., Richard Hichwa, Ph.D.
Funding Status: Project received a fundable priority score and percentile (16.6%) 11/97.
Project Status and Duration: Ongoing

Specific Aims
The specific aims of the project are to study the rCBF changes in brain systems mediating cognitive functions that are acutely impaired by smoking marijuana in occasional users, and to determine if long-term use of marijuana alters the acute response of these brain systems to marijuana. To accomplish these aims we will perform an interrelated series of 4 studies using PET with [15O]water.

Study 1. Time Course of rCBF Changes Following Smoking of Marijuana: This initial study will assess the time course of rCBF changes in the brain after smoking marijuana.

Studies 2. & 3. Acute Effects of Marijuana on Memory and Attention: Two studies will assess the effects of smoking marijuana on memory and attention in occasional users.

Study 4. Acute Effects of Marijuana on rCBF and Memory in Chronic Users: This study will assess the manner in which heavy, long-term use of marijuana alters the acute response of the brain to marijuana during performance of a memory task.

Background and Significance
Marijuana (cannabis sativa) is the most commonly used illegal drug in the United States (Preliminary Estimates From the 1994 National Household Survey on Drug Abuse 1995). Smoking of marijuana causes subjective effects including euphoria, depersonalization, altered time sense, lethargy, drowsiness, confusion, anxiety, and psychosis (Hollister 1988); as well as impairment on sensory, motor, and cognitive tasks (Block et al 1992). The major psychoactive substance in marijuana, delta-9-tetrahydrocannabinol (THC), is one of a class of natural and synthetic cannabinoids that appear to selectively bind to receptors in the hippocampus and cerebellum, as well as in outflow nuclei of the basal ganglia (Herkenham et al 1990). The few imaging studies that have assessed the effects of marijuana and THC on brain blood flow and metabolism in humans have all utilized a "resting" baseline condition in which subjects perform no explicit task and have produced conflicting results. The studies proposed here will control the subject's mental activities following the smoking of marijuana or placebo by engaging them in challenging tasks. We will systematically assess the changes caused by smoking marijuana versus placebo on rCBF and tasks requiring attention and memory. This will permit assessment of the changes in brain activity that are associated with impaired behavioral performance.

Hypotheses Study 1:
1. Global blood flow will increase following the smoking of both placebo and marijuana cigarettes, but task-related rCBF decreases in temporal lobe will be seen only following marijuana smoking.
2. Significantly increased rCBF will also be observed in the cerebellum and frontal lobes studies (Mathew and Wilson 1993; 1995; Volkow et al 1991, 1996c), and the cerebellar increase will be positively correlated with subjective ratings of intoxication (Volkow et al 1991).

Hypotheses Study 2 and 3:
The general hypothesis for the two studies is that smoking marijuana will attenuate the task-specific pattern of rCBF that is characteristic of normal cognitive function in occasional users. That is, we hypothesize that areas of activation will be smaller in spatial extent and lower in magnitude following marijuana, in conjunction with poorer cognitive performance.

Memory:
1. The rCBF activations in the placebo condition will be similar to our previous findings. That is, we expect both novel and familiar recall conditions to activate right frontal, left inferior temporal, anterior cingulate, thalamus, and cerebellum.
2. Activation in the regions noted in Hypothesis #1 will be smaller in extent and lower in magnitude in the marijuana session, in conjunction with poorer recall scores.
3. The effects of marijuana on rCBF will be significantly greater in the more difficult novel word list condition than in the practiced condition.

Attention:
1. We will replicate findings from our prior PET studies in the placebo session. That is, we expect that rCBF asymmetries will vary with attention, with greater flow in the superior temporal gyrus (STG) contralateral to the direction of attention.
2. Smoking marijuana will attenuate this attention-related asymmetry.
3. Smoking marijuana will decrease attention-specific changes in rCBF in the planum temporale and in the superior parietal lobe, but will not change rCBF in primary auditory cortex in Heschl's gyrus.

Hypotheses Study 4:
1. The acute rCBF response to marijuana in chronic users during the baseline task will replicate the pattern of cerebral metabolism change recently reported by Volkow et al (1996c) in marijuana abusers injected with THC. That is, we expect that chronic users will show rCBF increases in orbitofrontal cortex, prefrontal cortex and basal ganglia.
2. Task-specific patterns of rCBF change observed during memory task performance will show the same general pattern of attenuation as in occasional users (i.e., rCBF changes smaller in extent and lower in magnitude following marijuana versus placebo), but the attenuation will be significantly greater in the chronic user group.

Progress Report
We have completed PET imaging utilizing the protocol described in Study 1 above in four subjects and have recently completed our first group t-map analysis of the effects of smoking marijuana. To our knowledge, we are the first to investigate the effects of smoking marijuana on brain function using PET during the performance of a cognitive task. We found rCBF increases in a number of brain regions (e.g., frontal lobes, lateral cerebellum) that showed increased rCBF in previous imaging studies (Mathew and Wilson 1993; 1995; Volkow et al 1991; 1996c), which assessed the effects of marijuana on subjects during a resting state. Additionally, we found dramatic rCBF reductions following marijuana smoking in regions of the temporal lobe that we have found to be sensitive to attentional effects during the auditorily-presented cognitive task (O'Leary et al 1996 a, b; 1997a, submitted a). It seems possible that the reduction of task-related rCBF in temporal lobe cortices may be a direct neurobiological correlate of the acute cognitive impairment that is caused by smoking marijuana.

Research Plan:
Sample: Each study will utilize 20 subjects (10 males and 10 females). Three studies will use subjects who report using marijuana no more than 10 times a month. Study 4 will utilize chronic users of marijuana (subjects who report using marijuana 7 or more times weekly on average over the past two years). Scores on the Iowa Tests will be used to match the chronic users of marijuana in Study 4 as a group, with the scores of the subjects in the other studies.

Smoking in each study will take place while the subject is on the PET couch, at the time point in the study noted below. We found in our nicotine study that subjects can hold their own cigarettes even with both arterial and venous lines in place. Subjects will follow a paced smoking procedure to control the dose of THC (Block et al 1992). A custom-designed ventilation hood will be in place directly above the subject to vent the smoke. Marijuana or placebo cigarettes will be obtained from NIDA and will be stored and monitored by the pharmacy at the University of Iowa Hospitals and Clinics. The active cigarette will contain 20 mg THC, a dose which we have found to produce substantial, but not incapacitating behavioral effects (Block et al 1992); and which is representative of typical social use. Placebo cigarettes will contain inactive, cannabinoid-extracted marijuana with only trace amounts of THC.

Study 1:
Year 1. Time Course of rCBF Changes Following Smoking of Marijuana: A double-blind marijuana/placebo design will be used, with half of the subjects randomly assigned to a condition in which marijuana will be smoked following injection #3 and placebo smoked following injection #5, and half of the subjects smoking placebo and marijuana in the opposite order. The activation conditions for this study are: Sham, reaction time (RT) baseline; condition #1, RT baseline; conditions #2 through #8, attention task (dichotically-presented CVCs - attend left). Should rCBF not remain stable over the time course to be used in following studies, we will have subjects smoke a second marijuana cigarette in these studies.

Studies 2-4:
Each of the following studies will have essentially the same design. Studies 2 & 3 will differ in the activating tasks that are used with subjects who are occasional users of marijuana, whereas Study 4 will use the same memory tasks as Study 2, but will use subjects who are chronic users of marijuana. Each of the studies will involve two sessions at least one week apart with four conditions per session, and will be doubly-blinded, with order of placebo or marijuana counterbalanced. Each session will involve an initial sham injection followed by four conditions during which [15O]water will be used to measure rCBF. All studies will take place in the morning and will last approximately two hours. The sham and an initial condition will use a baseline task that is unique to each study, followed by smoking a standard marijuana or placebo cigarette. A second repetition of the baseline task will then take place followed by two activation conditions. This design permits within-subject comparisons of the effects of smoking marijuana vs placebo both within- and across- sessions. This design yields a number of possible comparisons of conditions that will be of interest in each study. All eight rCBF PET images (four from each session) will be co-registered with each individual's MR images, permitting within-subject subtractions both between- and within-sessions. Each session has the same baseline that is repeated prior to and following smoking of marijuana vs placebo. Since order of sessions will be counterbalanced, this will permit an analysis of variance design with one between subject variable (order of sessions), and two within-subject variables (pre vs post smoking of marijuana vs placebo.

Study 2.
Effects of Marijuana on rCBF and Memory Year 2 & 3. The activation conditions for this study, to be repeated on two occasions, are: Sham, counting baseline; #1, counting baseline; #2, counting baseline; #3, practiced list; #4, novel list. Smoking will occur prior to condition 2. This study will assess the effects of marijuana vs placebo on the retrieval of practiced and novel word lists. The verbal recall conditions will be similar to conditions we have used previously (Andreasen et al 1995e, f). We will, however, use an active baseline task rather than a resting baseline for reasons discussed above, and in Andreasen et al (1995d).

Study 3.
Effects of Marijuana on rCBF and Attention Year 3. The activation conditions for this study, to be repeated on two occasions, are: Sham, reaction time (RT) baseline #1, RT baseline, #2, RT baseline, #3, dichotically-presented CVCs - attend right #4, dichotically-presented CVCs - attend left. Conditions 2 and 3 will be counter-balanced across subjects. Smoking will occur prior to condition 2. This study will assess the effects of marijuana vs placebo on the ability to focus attention on relevant information and to filter out irrelevant information. The baseline and dichotic conditions to be used are identical to conditions we have used in our PET work with normal volunteers and patients (O'Leary et al 1996 a, b).

Study 4.
Acute Effects of Marijuana on rCBF in Chronic Users Years 4 & 5. This study will compare rCBF during marijuana challenge in a group of subjects who have used marijuana on a regular basis for a number of years to rCBF changes in subjects who are occasional users. Chronic use is defined as using marijuana on average 7 times or more weekly over the past two years and with a total duration of use of at least 4 years. The design of the study will be identical to Study 2 described above. The memory activation tasks involve verbal retrieval processes that our previous work has shown to be impaired in chronic users of marijuana. In addition to within-group analyses of the acute effects of marijuana we will compare the acute effects of smoking marijuana versus placebo on rCBF in the group of subjects who are chronic users of marijuana to rCBF changes in subjects who are occasional users.

Data Analysis:
The PET data will be analyzed utilizing the image analysis procedures described above and in the Image Analysis Core Unit. For Studies 2-4, within-subject analyses will utilize a subtraction image t-map approach to compare the same conditions during the marijuana and placebo sessions. Randomization analysis (Arndt et al 1996c) will be used for between-subject analyses comparing chronic users (Study 4) with occasional users of marijuana.

Back to top


PROJECT 6: COGNITIVE FUNCTIONING, BRAIN STRUCTURE AND METABOLIC ACTIVITY IN CHRONIC ALCOHOL ABUSERS
Principal Investigator: Karen M. Cocco, Ph.D.
Co-Investigators: Peter E. Nathan, Ph.D., Nancy C. Andreasen, M.D., Ph.D., Daniel S. O'Leary, Ph.D.
Project Status and Duration: Currently funded through the CIFRE mechanism

Specific Aims
1. To use PET imaging with [15O]water to assess rCBF changes in long-term alcohol abusers during the performance of attention and memory tasks that have been shown to be sensitive to prolonged alcohol abuse.
2. To use PET to assess rCBF changes associated with recovery of cognitive function following abstention from alcohol for several months.
3. To use MR imaging to assess structural brain changes associated with long-term alcohol abuse and to assess structural changes that may occur following abstention.
4. To assess changes in cognitive function associated with long-term alcohol abuse and abstention using an extensive neuropsychological test battery.

Background and Rationale
Substantial data document the impact of prolonged alcohol abuse on cognitive functioning. Approximately 50-70% of detoxified alcoholics experience deficits in cognitive functioning. Problem-solving abilities, the ability to make conceptual shifts, and short-term memory and attentional processes are most sensitive to the effects of prolonged, excessive ethanol consumption. Data also exist that suggest strongly that chronic consumption of alcohol is related to both structural and functional changes in the brain. Increases in the ventricular brain ratio (VBR) and increases in CSF in sulci and interhemispheric fissures were the most frequently reported indices of the brain damage resulting from prolonged, excessive drinking in early CT studies, and these findings have been confirmed in more recent MR imaging studies (Johnson et al 1986). Twin studies have demonstrated that these changes are a consequence of prolonged heavy drinking rather than pre-existing conditions. The modest PET data that have been reported confirm that alcoholic drinking affects both cortical and subcortical structures (Gilman et al 1990; Volkow et al 1992; Wik et al 1988). The medial frontal region of the cerebral cortex seems to be most affected (Gilman et al 1990; Melgaard et al 1990; Samson et al 1986), although a small area in the left parietal region also appears to be involved (Melgaard et al 1990; Volkow et al 1992; Wik et al 1988).

These initial investigations have not to date examined interrelationships among structural changes in the brain, changes in metabolic activity, and alcohol-induced deficits in neuropsychological functioning. Those studies that have looked simultaneously at indices of brain structure and function, like the investigation by Volkow et al (1992), have only done so at a single point in time, even though ample neuropsychological data suggest that substantial recovery of cognitive deficits occurs over time with abstinence. While there is modest evidence to suggest that recovery of functioning as measured by both neuropsychological test performance and structural and functional changes in the brain is likely to occur within the first 4 months of abstinence from alcohol, it is difficult to interpret the long-term implications of these findings because these assessments were all done at a single point in time, when alcohol-related changes, including the effects of withdrawal, on both brain structure and function would be expected. The interrelationships among changes in brain structure, underlying pathophysiological processes in the brain, and neuropsychological deficits as a function of the post-abstinence interval have not yet been examined since the only studies in which all three measures of dysfunction were employed did so on only a single occasion.

Hypotheses
1. Long-term abusers of alcohol will show significant cognitive deficits and significant differences from normal volunteers in rCBF during the performance of attention and memory tasks when assessed within a month from abstention.
2. Regional CBF will "normalize" in association with improved cognitive performance when the same individuals are assessed after several months of abstinence from alcohol.
3. MR imaging will show significantly increased ventricular and surface CSF in long-term abusers of alcohol when imaged within a month from abstention, but this atrophy will reverse to some extent following abstention.
4. There will be a significant relationship between the severity of cognitive deficits associated with long-term alcohol abuse and structural and functional brain measures.
Progress Report
This project has received a small grant from the CIFRE mechanism in order to permit collection of pilot data from 5 subjects. The data to be gathered in this pilot study will be used to establish the feasibility and some of the parameters of a substantially more extensive study of relationships among neuropsychological deficits and structural and functional changes in the brain both following prolonged alcohol abuse and after a several-month period of abstinence from alcohol in chronic alcohol abusers. Subsequent investigations will examine samples of chronic alcohol abusers who differ in age, chronicity of alcoholism, gender, family history and polysubstance. The project has been approved by the University of Iowa Human Subjects and Radiation Protection Committees and the first subject has been recruited for the project.

Research Plan
Patient Sample:
Subjects will be adult male inpatient volunteers participating in a substance abuse treatment program located in Iowa City (MECCA). All subjects will be between the ages of 21 and 35 years of age; strongly right-handed; without a past history of hospitalization for (DSM-IV) diagnoses of schizophrenia or major affective disorders; without a history of medical or neurological illness or trauma that would affect the central nervous system; without seizure disorder, related or unrelated to substance use; without a history of polysubstance abuse; without a family history of alcohol dependence; and at least 3 weeks but no longer than one month, abstinence from alcohol before the first MRI, PET, and neuropsychological measurements. Control subjects will be 13 adult male volunteers who previously participated in an independent investigation (Andreasen et al 1995e, f) using the same paradigm proposed herein.

Methods
The project will utilize PET with [15O]water to measure regional cerebral blood flow at two different points in time in 5 adult, male individuals currently in treatment for alcohol dependence. Two recall tasks will require the subject to remember stories, each of which contains 25 concepts. For a "practiced recall" condition, subjects will be trained to perfect recall of the 25 concepts and will subsequently be asked to recall as many concepts as possible during one PET imaging condition. For the "novel recall" condition, subjects will be exposed to a second story 1 minute prior to PET scanning and will be instructed to recall as much of the novel story as possible during the PET imaging condition. These two tasks have been chosen because the literature indicates that long-term memory remains relatively intact after a period of abusive drinking while new material is substantially more sensitive to the effects of alcohol abuse. Therefore, the practiced recall condition will act as a control condition for each subject. MRI will be used to reflect brain structure at the same two time points as the PET imaging. PET and MRI acquisition and analyses will utilize the techniques described above and in the Image Analyses Core Unit. Within-group analyses of changes between the two assessments will utilize randomization analysis (Arndt et al 1996c), as will between-group analyses comparing rCBF at each of the two time points in the alcohol abuse group to the control group.

Data Analysis:
The PET and MR imaging data will be analyzed using the procedures described above and in the Image Analysis Core Unit. Within-group analyses of PET data will utilize a subtraction image, t-map approach. Within-group changes in MR images will be analyzed by co-registration of each subjects images from the two occasions, subtracting the two image sets and computing an effect-size map for the group (Andreasen et al 1994k). Analyses comparing the chronic alcohol abuse group to controls will utilize randomization analysis (Arndt et al 1996c).

Back to Top

PROJECT 7: EFFECTS OF CIGARETTE SMOKING / NICOTINE EXPOSURE ON COGNITIVE PERFORMANCE AND REGIONAL BRAIN ACTIVITY AS ASSESSED BY FUNCTIONAL MRI IN PATIENTS WITH SCHIZOPHRENIA

Principal Investigator: Michael Flaum, M.D.
Co-Investigators: Daniel O'Leary, Ph.D., Diane Mosnik, M.S., James C. Ehrhardt, Ph.D., Vincent Magnotta, Ph.D., Stephan Arndt, Ph.D., Ted Cizadlo, B.S., Jane S. Paulsen, Ph.D.
Funding Status: This project is funded by a grant from the Nellie Ball Trust Research Fund to Dr. Flaum
Project Status and Duration: This project has recently begun - the grant is for 1 year from 10/97 - 9/98

Specific Aims
1. To quantify the effects of two types of nicotine exposure (cigarette smoking and nicotine patch) on cognitive performance among smokers and non-smokers with schizophrenia, and control smokers and non-smokers.
2. To estimate the location and magnitude of the effects of varying nicotine exposure on regional brain activity during performance of selected cognitive tasks.

Background and Significance:
The recent report (Freedman 1997) linking a gene at the site of the alpha7-nicotinic receptor to vulnerability for information processing deficits among patients with schizophrenia and their families, has increased interest concerning cigarette smoking in individuals with schizophrenia (Lohr 1992). Several lines of converging research are of interest. First, nicotine, at least when delivered in the form of cigarette smoking enhances performance on specific types of cognitive tasks among normal smokers (Wesnes 1983) (there is far less certainty regarding its effects on non-smokers); specifically, those of focused/vigilant attention, choice reaction time (RT) tasks, motor speed and dexterity, and certain memory tasks, including those tapping working memory. There is some specificity to the effects, in that other cognitive functions such as long term memory tasks, actually appear to be worsened by nicotine (Spilich 1992). There is overlap between the nicotine enhanced tasks and the types of cognitive deficits that have been implicated as potential primary abnormalities in schizophrenia (Braff 1993). Second, although the pharmacology of nicotine is complex, effects on dopaminergic systems, particularly in the meso-limbic structures are among the most prominent (Grenhoff 1988; 1989). Third, a series of studies over the past few years have shown a transient "normalizing effect" of cigarette smoking on a variety of characteristic deficits of schizophrenia including sensory gating (Adler 1992) and smooth pursuit eye movements (Klein 1991). Combined with the extremely high rate of smoking in this population these data suggest that schizophrenics are more prone to use nicotine in an effort to self-medicate an underlying attentional and information processing deficit.

Hypotheses:
1. Performance on each of the cognitive tasks will improve following cigarette smoking relative to both the nicotine patch and abstinence conditions across all subject groups.
2. Regional brain activity, as assessed by fMRI, will increase as a function of task performance as evidenced by both the number of pixels surpassing a prespecified threshold of activation, and the average intensity of activated pixels.
3. Regional brain activity (combining hypotheses 1 and 2) will increase as a function of nicotine exposure, with recent cigarette smoking yielding greater activations than either abstinence or the nicotine patch condition.
4. The effect of nicotine on cognitive performance and regional brain activity will be greater among the smokers than non-smokers, and among the schizophrenic smokers than control smokers. (i.e., interactions are expected between smoking status, diagnosis and nicotine exposure on cognitive performance)

Preliminary Studies
This project required solving a number of novel technical problems since we wished to measure rCBF and levels of nicotine and several other substances in plasma both before and immediately after smoking a tobacco cigarette following a period of abstinence. The study required the construction of a ventilation system within the PET suite, to eliminate the tobacco smoke. The study also required development of protocols for the rapid assessment of levels of carbon dioxide, carbon monoxide, nicotine and metabolites; as well as software permitting the voxel by voxel correlation of circulating levels of substances and rCBF. The equipment, software and experience was later used in our pilot studies of the effects the smoking marijuana cigarettes.

An initial study of the effects of smoking tobacco cigarettes has been completed in five healthy young men, all of whom were regular cigarette smokers (Flaum et al submitted a). The conditions across each of these ten scans were identical with one critical exception: Immediately prior to the third and eighth scan, the subjects smoked a full cigarette of their usual brand. All subjects abstained from smoking for 12 hours prior to the first scan. Subjects were engaged in the same attentional task (a dichotic listening paradigm) during each of the ten scans. Nicotine blood levels were obtained just prior to each scan. When whole brain blood flow was examined, no consistent or significant effect of smoking was demonstrated across subjects. However, regionally specific changes in flow as a function of nicotine levels were found. Specifically, blood flow in the occipital cortex and right medial temporal/insular cortex was positively correlated with plasma nicotine levels over time. Interestingly, the temporal/insular cortical areas in which flow increased as a function of nicotine level corresponds to aspects of cortex presumed to subserve auditory association function. Further, this cortical area also corresponded to that demonstrated to be "activated" by a dichotic listening task in a previous study of normal controls in our laboratory. These preliminary data suggest that cigarette smoking does not acutely cause a consistent increase or decrease in whole brain or regional cerebral blood flow. However, plasma nicotine levels do affect blood flow in a regionally specific manner, as blood flow in both the visual and auditory association cortices vary as a function of plasma nicotine levels. These data are consistent with neuropsychological findings which suggest that nicotine may serve a "cognitive enhancing" function. The reviewers of this project noted that task performance might be expected to improve following smoking, since subjects are likely to be addicted to nicotine and become psychologically disrupted if they do not smoke on schedule.

Methods
In this project, we are measuring performance on selected neuropsychological tasks (chosen from those found to be most sensitive to nicotine exposure among normal smokers) in four subject groups: those with schizophrenia and normal controls, each divided into regular smokers and non-smokers. Measures are obtained during multiple sessions for all groups, varying the duration since last nicotine exposure (via cigarette smoking and nicotine transdermal patches among the smoking groups, and transdermal patches only among the non-smokers). Functional magnetic resonance imaging (fMRI) is used to identify the brain regions associated with performance on selected cognitive tasks, and to quantify the effects of varying nicotine exposure on regional brain activity during performance on these tasks. We are extending our pilot work in a number of ways. First, we are obtaining data on non-smokers (both patients and controls) as well as smokers. If we are seeking a "normalizing" phenomenon, it is critical to have a sense of what "normal" really is; normal smokers may be a poor proxy for this. Existing data are available regarding the cognitive performance aspect of this, but not for the physiological correlates (e.g., brain imaging). Second, we are including a nicotine patch condition, in addition to a recent smoking vs. abstinence condition. This allows us to control the dose of nicotine across subjects and within subjects over time. Third, we are using fMRI rather than PET. Reasons for this include the ability to repeat subjects on multiple occasions as well as to lower the expense.

We are studying four groups of 10 subjects each: patients and controls, smokers and non-smokers (smokers defined as those regularly smoking > 1 pack/ day for at least the past six months). The patient group is selected primarily from those participating in ongoing MH-CRC research protocols which require a three week psychoactive medication "washout" period. In order to minimize confounding effects of treatment, patients are being studied while off medications for at least two weeks. All patients in the study will have a diagnosis of schizophrenia or schizoafffective disorder, by DSM criteria. Controls are selected from the community on the basis of equivalence with the patient group in terms of: age, sex, parental education and parental socioeconomic status. Sample size choice was informed by effect sizes from numerous imaging studies in our lab and others (Andreasen 1996j).

Timing of cognitive testing and fMRI sessions:
Each subject undergoes three separate sessions of neuropsychological assessment, and three corresponding fMRI sessions, varying only the timing and type of nicotine exposure. The cognitive testing and fMRI sessions are conducted on the same day in the following three conditions:
1) After 24 hours of abstinence from cigarette smoking or any other form of nicotine
2) After at least a 24 hour period of smoking "ad lib" (own brand, self-dosing), and smoking a cigarette just prior to getting in the scanner (or just prior to beginning testing)
3) While wearing a nicotine transdermal patch (dosed to be equivalent to usual smoking dose) for the previous three days, while abstaining from cigarette smoking.

Cognitive Tasks:
Tasks have been chosen that have been shown to be sensitive to nicotine effects among normals and shown to be resistant to a learning or practice effect. Examples include: motor tasks, e.g., "Purdue Pegboard" (manual dexterity), finger oscillation task (simple motor speed), and a complex motor sequencing task; vigilant attention tasks, e.g., continuous performance task; reaction time tasks, e.g., the "Flanker" task; and working memory tasks e.g., the "1back / 2back and the "Sternberg" tasks.

fMRI Methods:
Not all of the cognitive tasks are adaptable to the fMRI environment (e.g., vigilant attention tasks); however we have adapted two of the motor tasks, one reaction time (RT) task (the Flanker), and one working memory task ("1-back / 2back") for the fMRI environment. The fMRI sessions typically consist of 3 - 4 "runs" of image acquisition, (a motor, memory and RT task + one repeat) each lasting four minutes. The order of the four runs is constant across sessions within subject, but is varied across subjects in order to partially account for order effects. All subjects are pre-exposed in our "mock scanner" which simulates the visual, auditory and posture/positioning of the MR scanner. Nicotine and cotinine levels are drawn at the conclusion of each fMRI session. Tidal-breath carbon monoxide (CO) is measured prior to each fMRI run. Behavioral data and reaction times are collected during the fMRI sessions (via a hand held manipulandum), as well as outside of the fMRI setting. Heart rate and blood pressur