{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp04_retrieval_first_paired20_20260802","cell_id":"invariant_mode_decomposition_design__psychology","hypothesis_id":"H2","search_queries":["exposure therapy adaptive difficulty trial by trial fear safety behavior approach avoidance dynamical systems","virtual reality exposure therapy closed loop physiological adaptive difficulty patent","exposure therapy computational model dynamical systems fear avoidance safety behavior trial","inhibitory learning exposure therapy expectancy violation fear reduction pacing hierarchy guidelines","state space model exposure therapy trial level fear avoidance arousal latent dynamics","system identification exposure therapy adaptive control multivariate physiological behavioral response","\"exposure therapy\" \"dynamical systems\" avoidance arousal","patent adaptive exposure therapy multivariate physiological behavioral model difficulty"],"sources":[{"source_id":"C1","title":"US11523773B2 — Biofeedback for therapy in virtual and augmented reality","publisher":"United States Patent and Trademark Office / Google Patents","url":"https://patents.google.com/patent/US11523773B2","source_class":"GOVERNMENT_OR_REGULATOR","claims_supported":["The patent describes collecting motion and biometric data, supplying them to a learning system, and adjusting a training protocol from the combined feedback.","For exposure therapy specifically, the system can scale exposure intensity and duration based on biometric feedback.","This closely anticipates closed-loop, response-dependent exposure pacing, but the disclosed examples use thresholds or a learning system rather than explicit modal stability and sensitivity analysis."]},{"source_id":"C2","title":"US20200302825A1 — Automated selection and titration of sensory stimuli to induce a target pattern of autonomic nervous system activity","publisher":"United States Patent and Trademark Office / Google Patents","url":"https://patents.google.com/patent/US20200302825A1/en","source_class":"GOVERNMENT_OR_REGULATOR","claims_supported":["The application describes automatic quantitative titration of exposure severity using real-time physiological signals.","It selects or changes sensory content to track a desired temporal pattern of autonomic activity and discusses sequential machine-learning models operating across time.","It identifies both overly rapid and overly slow exposure progression as treatment problems, strongly supporting the nominated pacing problem while chiefly optimizing physiological or anxiety targets."]},{"source_id":"C3","title":"Automated Personalized Exposure Therapy Based on Physiological Measures Using Experience-Driven Procedural Content Generation","publisher":"AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment","url":"https://ojs.aaai.org/index.php/AIIDE/article/view/18914","source_class":"PRIMARY_RESEARCH","claims_supported":["The proposed VRET framework predicts individual experience from physiological sensors and machine learning.","It automatically adapts exposure parameters from the subject's physiological response.","The 2021 document is a doctoral-consortium report describing intended human studies, so it establishes the approach more strongly than its clinical efficacy."]},{"source_id":"C4","title":"Personalizing Exposure Therapy via Reinforcement Learning","publisher":"arXiv","url":"https://arxiv.org/abs/2504.14095","source_class":"PRIMARY_RESEARCH","claims_supported":["The system uses reinforcement learning and physiological measures to adjust virtual-spider attributes during arachnophobia exposure.","A human-subject comparison evaluates the adaptive controller against a rules-based method.","Its controller seeks therapist-defined anxiety levels, rather than estimating eigenmodes of a joint fear, approach, avoidance, arousal, and safety-behavior transition."]},{"source_id":"C5","title":"Tutorial for Using Control Systems Engineering to Optimize Adaptive Mobile Health Interventions","publisher":"Journal of Medical Internet Research","url":"https://www.jmir.org/2018/6/e214/","source_class":"AUTHORITATIVE_SECONDARY","claims_supported":["The tutorial presents within-person system identification as dynamic theory testing for adaptive behavioral interventions.","It recommends deliberately varying intervention options to excite the system, estimating and validating idiographic dynamical models, and using those models in controllers.","This supplies most of the neighboring methodological workflow behind graded-trial excitation and model-based pacing, although it is not an exposure-therapy modal controller."]},{"source_id":"C6","title":"Real-Time Applications of Biophysiological Markers in Virtual-Reality Exposure Therapy: A Systematic Review","publisher":"BioMedInformatics / EBSCOhost","url":"https://openurl.ebsco.com/contentitem/doi%3A10.3390/biomedinformatics5030048?id=ebsco%3Adoi%3A10.3390%2Fbiomedinformatics5030048&sid=ebsco%3Aplink%3Acrawler","source_class":"AUTHORITATIVE_SECONDARY","claims_supported":["The review found fifteen VRET studies combining physiological monitoring with adaptive features.","Ten used biofeedback from single or multimodal physiological measures, while others used dynamic categorization or machine learning to alter scenarios or threat.","It characterizes adaptive VRET as technically feasible but supported by heterogeneous, generally small studies."]},{"source_id":"C7","title":"Treatment of avoidance behavior as an adjunct to exposure therapy: Insights from modern learning theory","publisher":"Behaviour Research and Therapy / PubMed","url":"https://pubmed.ncbi.nlm.nih.gov/28477845/","source_class":"AUTHORITATIVE_SECONDARY","claims_supported":["The review identifies pathological avoidance and safety behavior as distinct exposure-treatment targets.","It reports that avoidance can persist after treatment and that merely retaining an avoidance option can renew fear after otherwise successful extinction.","This supports the concern that fear reduction alone can misrepresent clinically important response dynamics."]}],"proximity":"SUBSTANTIAL_COLLISION","closest_analogues":[{"name":"Biofeedback-controlled VR exposure titration","similarity":"Very high workflow overlap: repeated exposure responses are sensed, combined physiological and behavioral or motion data can enter a learning system, and subsequent exposure intensity, duration, or difficulty is adjusted automatically.","remaining_difference":"The located patent uses learned outputs or physiological thresholds and does not disclose a trial-to-trial state-transition eigendecomposition with explicit dominant-mode stability and sensitivity classes.","source_ids":["C1","C2"]},{"name":"Physiology-driven reinforcement-learning VRET","similarity":"Very high functional overlap: an individualized closed-loop controller changes exposure content in real time from physiological responses and has been compared with a rules-based controller.","remaining_difference":"The optimization target is a desired anxiety level; approach, avoidance, safety behavior, and arousal are not jointly decomposed into invariant modes that determine advance, hold, or regression.","source_ids":["C3","C4","C6"]},{"name":"Idiographic system identification for adaptive behavioral interventions","similarity":"High mechanism-level overlap: interventions deliberately excite an individual system, sequential responses identify a dynamical model, and a model-based controller assigns later intervention doses.","remaining_difference":"The guidance is domain-general and does not demonstrate exposure-trial pacing based on eigenmode stability or outcome sensitivity.","source_ids":["C5"]},{"name":"Avoidance- and safety-behavior-informed exposure practice","similarity":"Moderate clinical overlap: it rejects fear reduction as a complete account and treats avoidance and safety behavior as independently consequential treatment processes.","remaining_difference":"It supplies clinical targets rather than a multivariate modal estimator or automated difficulty-selection rule.","source_ids":["C7"]}],"overlapping_components":["Closed-loop selection of the next exposure's intensity, duration, difficulty, or content","Patient-specific feedback from repeated exposure responses","Use of multiple physiological and behavioral or motion measurements","Within-person temporal modeling and system identification","Deliberate variation of intervention dose to excite and learn individual dynamics","Automated advance or regression rather than a fixed hierarchy","Recognition that overly rapid escalation and overly slow progression can both be harmful","Recognition that avoidance and safety behavior can remain clinically important despite fear reduction"],"remaining_contrastive_claim":"Between exposure trials, difficulty is selected by estimating a multivariate transition operator over fear, approach, avoidance, arousal, and safety behavior, decomposing it into dominant modes, and using each mode's stability class and intervention sensitivity—rather than a scalar anxiety target, physiological threshold, or undifferentiated learned policy—to advance, hold, or regress.","claim_falsifier":"A pre-existing public paper, product specification, clinical protocol, or patent would falsify the contrastive claim if it jointly disclosed trial-level exposure-state estimation across behavioral and arousal variables, eigenmode or equivalent invariant-mode extraction, explicit stability and intervention-sensitivity classification, and use of those classifications to choose the next exposure difficulty.","problem_support":"STRONG","recommendation":"RESEARCH","world_novelty_boundary":"This eight-query, seven-source ordinary-web review establishes substantial prior art for adaptive exposure pacing, multimodal biosensing, temporal machine learning, and idiographic model-based behavioral control, but it found no source containing the full modal decision rule. That is a bounded-search distinction, not a worldwide novelty finding; a full patent-family and claims search, bibliographic-database search, non-English search, product-documentation review, and terminology search for state-space eigenanalysis, latent dynamical systems, Koopman modes, and system-identification controllers in exposure therapy could erase it."}