STRUCTURAL AND FUNCTIONAL NEUROIMAGING FINDINGS IN THE BSNIP PSYCHOSIS CONSORTIUM

Abstracts of the 4th Biennial Schizophrenia International Research Conference / Schizophrenia Research 153, Supplement 1 (2014) S1–S384

for several cluster sizes searched (K=2 to 9), but the FNC window states of patients with SZ were more commonly assigned to a particular state of weak connectivity among subsystems whereas HC switched between different FNC states more often. Conclusions: Dynamic FNC analysis suggests that patients with SZ tend to linger in a state of “weak” and relatively “rigid” connectivity among different sub-networks. In contrast HC are probably faster in recruiting necessary resources as task demands change by dynamically switching between different FNC states while varying connectivity among sub-networks. Furthermore, patients with SZ showed weaker connectivity among sensory systems and hyper-connectivity between thalamus and sensory systems in different FNC states.

GUIDING NEUROMODULATION WITH NEUROIMAGING Kelvin Lim 1,2 1 University of Minnesota; 2 Minneapolis VAHCS The pace of medication development for psychotic disorders has slowed. In order to continue to improve treatment for patients suffering from psychotic disorders, innovative approaches need to be developed. Nonpharmacological interventions such as cognitive based approaches have shown promise for the management of a variety of symptoms observed in psychotic disorders. These cognitive based approaches depend on learning (brain plasticity), which can be impaired in psychotic disorders. Noninvasive neuromodulation methods can both probe and modulate brain plasticity and may be useful to enhance the effectiveness of existing treatments. It is likely that there is interindividual variation in brain network abnormalities as measured with current neuroimaging approaches. In order to be successful, neuromodulation interventions targeting putative brain network abnormalities will need to take this variation into account. Neuroimaging may be an important tool to identify neuromodulation targets.

NEW NEUROSCIENCE BASED COGNITIVE PARADIGMS FOR BIOMARKER RESEARCH IN SCHIZOPHRENIA Deanna M. Barch 1 , C.S. Carter, J. Gold 2 , M. Harms, A. MacDonald, A. Poppe, J.D. Ragland, S. Silverstein, M.E. Strauss 1 Washington University in St. Louis; 2 University of Maryland School of Medicine, Department of Psychiatry, Maryland Psychiatric Research Center (MPRC), Baltimore, Maryland Background: Cognitive neuroscience has seen an explosion of technical advances and new knowledge regarding the neural basis of cognition. This presentation will overview of the results of the Cognitive Neuroscience Task Reliability & Clinical Applications (CNTRACs) Consortium work developing imaging biomarkers for three of the cognitive tasks developed by the consortium: Goal Maintenance (Dot Probe Expectancy Task; DPX), Episodic Memory (Relational and Item Specific Encoding Task; RISE), and Visual Integration (Jittered Orientation Visual Integration Task: JOVI). Methods: The CNTRaCs Consortium conducted a five-site imaging study, with a demographically matched sample of 55 patients with schizophrenia and 50 healthy controls. Participants completed a behavioral practice session within one week of their baseline scan, and then completed a baseline scan session followed by a retest scan within 28 days. Three of the scanners were Siemen’s Tim TRIOs using a 12 channel headcoil, one was a Siemen’s Allegra using an 8 channel headcoil, and one was a Phillips Achieva using a 12 channel head coil. Each imaging session including high resolution T1 imaging based on ADNI sequences, a T2 acquisition (to aid in co-registration), 2 sets of field maps, and 11 BOLD runs using an EPI sequence. Participants completed 4 runs of the DPX, 3 runs of the JOVI, and 4 runs for the RISE (1 during encoding, 2 during item recognition and 1 during relational recognition). Order of task administration was counterbalanced across participants and all participants took a break half way through the session. fMRI data were pre-processed using standard procedures with the FMRI Expert Analysis Tool (FEAT) in the FMRIB Software Library (FSL version 4.1). Statistical analysis was performed using the general linear model in FEAT, with research site added as a co-variate. Results: For the RISE, controls had greater activation than patients in medial temporal lobe regions (right hippocampus and parahippocampal

S13

gyrus) when successfully retrieving objects that had undergone relational encoding. In addition, when participants made errors following relational encoding controls showed greater activation than patients in a set of bilateral prefrontal regions (inferior and middle frontal gyrus, anterior cingulate gyrus) often associated with error detection. Conversely, in the item encoding condition, there were no group differences observed in medial temporal or prefrontal activation during either successful or unsuccessful item recognition. For the DPX, controls had greater activation than patients in dorsolateral prefrontal cortex in the comparison of cues signifying greater versus lesser need for goal maintenance. For the JOVI, all participants showed linear increases in activation in visual, frontal and parietal regions as demands on visual integration increased and patients showed altered activation in superior parietal and inferior frontal regions. There were few site differences that interacted with group, and minimal differences in QA indices across sites. Analyses of test-retest reliability are ongoing. Discussion: These results illustrate the successful development and implementation of imaging biomarkers paradigms of tasks developed as part of the CNTRaCs Consortium, and provide evidence for the validity of these paradigms as measures of the neural systems associated with specific cognitive processes. These results also illustrate the sensitivity of these measures to identifying neural changes associated with cognition impairment in schizophrenia and pave the way for their use as outcome and predictor measures in studies focused on evaluating treatments to enhance cognitive function in psychosis.

STRUCTURAL AND FUNCTIONAL NEUROIMAGING FINDINGS IN THE BSNIP PSYCHOSIS CONSORTIUM Godfrey D. Pearlson 1,2 , Carol Tamminga 3 , Matcheri Keshavan 4 , Gunvant Thaker 3 , Brett Clementz 5 , John Sweeney 3 , Shashwath Meda 6 , Sagin Khadka 6 1 Olin Neuropsychiatry Research Center and Yale University; 2 Institute of Living; 3 Department of Psychiatry, University of Texas Southwestern Medical Center; 4 Department of Psychiatry, Harvard Medical School; 5 Department of Psychology, BioImaging Research Center, University of Georgia; 6 Olin Neuropsychiatry Research Center, Hartford Hospital/Institute of Living Background: increasing evidence suggests that brain-based biological marker data tend to cross traditional psychiatric diagnostic boundaries. Clinical symptoms used in psychiatric diagnosis tend to be diagnostically non-specific, and risk genes discovered to date are similarly broad with regard to illness risk. Given the above, the BSNIP 5-site endophenotype study quantified multiple endophenotype candidates, derived from cognitive, oculomotor, electrophysiological and neuroimaging domains across DSM-diagnosed schizophrenia (SZ), schizo-affective (SA) and psychotic bipolar (PBP) patients and their first-degree relatives and controls. For neuroimaging, we explored whether abnormalities were more associated with classic psychiatric syndromes/disease diagnoses or alternatively with symptom dimensions, i.e. “psychosis”, and crossed traditional diagnostic boundaries. We determined if abnormalities were heritable and if so, were they found in all first-degree relatives or only the subset with psychotic spectrum diagnoses. Methods: We report here data on 917: SZ, N=149; SA, N=90; PBP, N=115; Healthy controls (HC) N=200; and patients’ relatives (SZ-R, N=134; SA-R, N=100; and PBP-R, N=129). Participants underwent imaging at 3-T that included structural (s)MRI, functional (f)MRI during a 5 min. resting period at all sites and DTI at 2 sites (subject N=513). sMRI data were processed using (a) VBM and (b) Freesurfer, DTI with tract-based spatial statistics to define fiber tracts and voxel-based analysis on skeletonized data using threshold-free cluster enhancement to correct for multiple comparisons. fMRI used the Group-ICA toolbox deriving within-and between-component connectivity measures and ALFF analyses implemented in Matlab. Statistical analyses were done blind to diagnosis with research site as a covariate. Results: For VBM, widespread cortical and subcortical gray matter (GM) volume reductions were related with psychosis across diagnoses, including relatives with cluster-A features. Using DSM, there were progressive GM deficits from PBP to SA to SZ. With DTI, patients across diagnoses shared white matter (WM) FA deficits, most consistently within corpus callosum genu and body: SC-R displayed similar WM deficits while PBP-R resembled controls. Again, cluster-A relatives resembled patients, so that corpus callo-

S14

Abstracts of the 4th Biennial Schizophrenia International Research Conference / Schizophrenia Research 153, Supplement 1 (2014) S1–S384

sum FA abnormalities were strongly associated with the clinical psychosis state and trait. For fMRI ALFF, SZ and SA shared marked, widespread deficits that were seen to a lesser extent in PBP; none of these were heritable, with all relatives resembling controls. With regard to resting network connectivity, 7 networks revealed abnormalities. Two of these were unique to schizophrenia probands (fronto-temporal/and cerebellum/midbrain), other abnormalities including the posterior default mode network, and were shared by probands. Thus, SZ, SA, PBP probands and their relatives shared both unique and overlapping brain connectivity abnormalities, some of which were seen in unaffected relatives. Discussion: these results reveal that neuroimaging data can be integrated across multiple sites without introducing site-by-diagnosis confounds. Different imaging measures clearly display different properties; in general, simpler measures (sMRI, DTI) tended to be more heritable. Many measures emphasize the predominant effect of these biological, quantitative abnormalities tracking more reliably with the symptom dimension of psychosis than with DSM diagnostic categories. While some imaging measures such as DTI-derived FA were strongly heritable, others such as fMRI functional connectivity were more variable.

Symposium SHOULD WE CONTINUE TO DO PLACEBO-CONTROLLED MEDICATION TRIALS IN SCHIZOPHRENIA? AN ETHICO-CLINICAL DEBATE Chairperson: Anthony S. David Discussant: William T. Carpenter Sunday, 6 April 2014 4:15 PM – 6:15 PM Overall Abstract: Therapeutic trials of medication in schizophrenia research, especially those sponsored by the pharmaceutical industry, often contain a placebo arm. Given that there are available several proven therapeutic agents for the treatment of this disorder, why is this necessary? Should not all trial compare a new trial gent with an established, efficacious one? In this symposium, we will debate the clinical, research, regulatory and ethical context behind placebo-controlled trials in schizophrenia therapeutics. This will be chaired by Anthony David, who is currently co-chair of the SIRS Ethical Committee. Each contributor has considerable experience of conducting such trials in various capacities. The symposium will start with Professor Wolfgang Fleischhacker a well-known clinical researcher from Austria, who will begin with an overview of the practical and methodological concerns which face researchers carrying out such trials. From his experience working with different companies as well as on academic studies, he will draw attention to the fact that methodological issues impinge on ethical ones and vice versa. After all a methodologically flawed study is arguably an unethical one and a ethically perfect study which is impossible to carry out in the real world will not benefit anyone. Next we will hear from Dr Luca Pani, a psychiatrist and representative from Italian Medicines Agency (AIFA)’s General Directorate, part of the European regulatory framework. He will explain how current guidelines allow for the safe use of placebos in clinical trials provided certain rules are adhered to. A view from the pharmaceutical industry will be provided by Dr Rob Conley who currently holds the position of distinguished scholar employed by Eli Lilly but has a long background in schizophrenia research in non-commercial settings. He will discuss the strengths and weaknesses of placebo controlled trials and will highlight some of the necessary safeguards for doing ethical research in this area. The issues of autonomy versus paternalism; informed consent and capacity to make decisions will be elaborated. The foregoing will set the stage for Paul Appelbaum, a US clinical psychiatrist and professional ethicist. Starting with the Declaration of Helsinki, he will go through how risk and harms can be balanced to enable useful, ethical research to proceed. Finally Will Carpenter will discuss all the contributions. Dr Carpenter has written on many of these topics and has unrivalled experience in the field of biological and social research in schizophrenia. He will broaden the consideration from placebos to the more general challenge of off-medication research. The symposium will attempt to explore and resolve the tension that sometimes develops between: the requirements of regulators to ensure that drugs may only be licenced after rigorous efficacy and safety checks; the aims of the pharma industry to produce novel agents and ensure profits for their shareholders and for the company to reinvest in

further discovery; the objectives of researchers to ensure that trial design and methodology succeeds in answering the questions they have posed; the aspirations of clinicians who are seeking ever better treatments for their patients. It is hoped that the symposium will lead to recommendations that may be considered by the SIRS Ethical Committee for dissemination.

THE USE OF PLACEBO IN RANDOMIZED CONTROLLED TRIALS IN SCHIZOPHRENIA: A REGULATOR’S VIEW Luca Pani Segreteria Tecnica Direzione Generale Agenzia Italiana del Farmaco The use of placebo represents a major challenge for CNS drug development. In the field of schizophrenia, in recent trials, the difference in efficacy between active treatments and placebo has tended to be smaller than the differences seen in the past. This has contributed to the increasing failure of registration trials and has raised ethical concerns. Informative data in clinical trials are those defined by a negative placebo response and a positive drug response. High placebo response rates are inversely related to the signals of efficacy thus reducing the efficiency of a trial. From a regulatory point of view, assay sensitivity cannot be guaranteed even in well designed and conducted trials if a placebo arm is not included. The latest version of the “Guideline on clinical investigation of medicinal products, including depot preparations in the treatment of schizophrenia” (EMA, 2012), states that predefined escape criteria, rescue medication and stopping rules, as well as a stringent follow up should be applied to safeguard patients’ safety in placebo controlled trials. Provided that, the benefits of using a placebo arm will generally override any ethical reservations in short term controlled efficacy trials. A placebo arm can still be possible and appropriate in long term studies if a randomized withdrawal design is applied to demonstrate maintenance of effect. On the other hand, recent work has shown that many factors could be identified affecting the level of placebo response, including both patients characteristics and trial design factors. New experimental designs, which control these factors, are being tested to verify if the likelihood to detect antipsychotic effect could be increased. A common effort is therefore needed among all stakeholders, such as patients, regulators, industry and research centers with Health Care Providers, which are involved in the drug development process and strategy to collaborate to optimize trial design in order to increase its efficiency, in terms of reducing the number of patients needed and reducing the placebo response rate with clear positive ethical implications.

THE ETHICS OF PLACEBO-CONTROLLED TRIALS IN SCHIZOPHRENIA: A VIEW INFORMED BY WORKING WITH INDUSTRY Robert R. Conley 1,2 Eli Lilly and Co; 2 Adjunct Professor, University of Maryland School of Medicine

1

People with schizophrenia present a challenge in the consideration of placebo controlled trials. Although there are commonly accepted standards of care in regard to medication worldwide it is recognized that the majority of people with this syndrome are not well treated by these current medication, which have serious problems with safety and efficacy. When there are no established effective interventions for the treatment of a disease, the use of a placebo control is not controversial. However, when an established effective intervention exists, the ethical acceptability of using a placebo control is sometimes challenged. If placebo control presents a serious risk, it is unethical. However, when the use of placebo control can reasonably be expected to result in only temporary or minor discomfort, it can be ethical. In addition to a low probability of morbidity, the use of a placebo control may be justified in situations in which existing treatments are minimally effective or have serious safety issues. Placebo-controlled trials are one of the most reliable ways to demonstrate the safety and efficacy of an investigational intervention because they provide a valid baseline against which the intervention can be compared. Placebo-controlled trials therefore have a great ability to distinguish between effective and ineffective treatment. This is crucial in conditions where treatments are not fully effective. Active-controlled trials are less sensitive and usually require a larger sample size than placebo-controlled trials. Therefore, a safety advantage of placebo-controlled trials is that they expose fewer