Resumen de Contributions to the development and validation of fMRI-based biomarkers for drug discovery in social anxiety disorder
This document presents the theoretical background and experimental work made to develop and validate a set of eperiments based on functional magnetic resonance imaging (fMRl). These eperiments are aimed to demonstrate that fMRl can be a valuable tool to objectiüze drug treatment response in Social Anxie§ Disorder (SAD) patients. Functional MRI is a non-invasive imag¡ng technique which proüdes localized indicators of brain actiüty, The analysis of fMRl data has recentlyfacilitated neuroscience to make a leap forward in the understanding of the human brain. Psychiatricclinical research,however,hasn'tfullyembracedyetthepotential of fMRI. lnparallel,thesocietal costs of new psychiatric drug discoveryare reaching unbearable limits. lt has been hypothesized thatthe addition of fMRl in clinical trials of pharm acologic treatments of SAD can proúde new biomarkers of treatm ent response which, in the future, s hall reduce the temporal and economicburden of drug discovery. Five studies are presented in this dissertation in an evolving path towards the validation of the hypothesis.
ln study 1, a widelyvalidated state-of-the-art emotional face processing paradigm was piloted in a non-clinical sample. Taskrelated activations were in line with the findings preüously reported in the literature. Howeler, the results of eperiment did not show a correlation with symptom severity. An additional eploratory psychophysiological interaction analysis revealed that the relationship between two emotion-processing areas had a significant correlation with SAD synptom severity. This emphasized the potential value of studies based on functional connectivi§forourobjectives.
Study 2 explored the reproducibilityof connectivityanalysis of fMRl data. To do so, a brain networkwas selected and explored with lndependentComponentAnalysis (lCA) on data collected from three categorically differentiated paradigms:Aresting task, a moral dilemma task and in a cognitive-challenging Stroop task. The selected network was systematically identified in the three cases, exempliñ/¡ng the robustness of the technique in three extreme cases.
Study 3 eplores the sens itiü§ of ICA by analyzing resting-state data acquired before and after an eperimental induction of sad mood. Multiple regions reflected changes in their intranetwork connectiütyafter sad mood induction. Results were validated using a split-half re-analysis and confirmatoryseed-based functional connectiü§measurements.These results supporltheideathatspatial lCAof fMRl is not only reliable, but also a sensitive paradigm.
Study4 presents a novel SAD symptom-provoking paradigm thatwas validated on SAD patients and controls.The analysis of this pilot revealed a striking non-linear relationship between task activation and social anietyscores. This non-linear relationship is compatible with some of the dircrgences found in literature regarding the alteration of emotional regulation brain areas in SAD.
Study5 presents the results of a small placebo-controlled clinical trial using a common treatmentforSAD (paroxetine) in SAD patients. Subjects underwentthe emotional face processing, the syrnptom'provoking and resting state tasks that were administered and analyzed following the experience obtained in studies 1 to 4.
The selected fMRl paradigms and analysis methods showed significant sensitiü§to the effects of paroxetine treatment on SAD. Treatment effects were identified in areas related to the processing of fear stress and aniety, which are known to be altered in SAD. Remarkably, ICA rerealed sensitiü§to pharmacologically-induced clinical improrement in the same areas and direction than the syrnptom-provocation task.
Along with the evidences reported in the literature reüew, the methods and res ults obtained throughout this dissertation produce a proof of concept on the usage of fMRl as a biomarkerfor SAD pharmacologic research.
