OGEN Analytics

Research Use Only. All metrics in this platform are exploratory and have not been validated as medical devices or clinical diagnostic tools. OGEN does not replace clinical assessment. All statistical comparisons are labeled "exploratory analysis" and must be interpreted by qualified clinicians.

לשימוש מחקרי בלבד. כל המדדים בפלטפורמה זו הם אקספלורטוריים ולא אושרו כמכשירים רפואיים או כלים אבחוניים קליניים. OGEN אינה מחליפה הערכה קלינית. כל ההשוואות הסטטיסטיות מסומנות "ניתוח חיוני" ויש לפרשן על ידי קלינאים מוסמכים.

Physiological Signals Metric Definitions Statistical Methods Algorithm Notes References

Physiological Signals

English

OGEN integrates two primary sensor modalities:

Verity Sense (Polar) — PPG: worn on the upper arm; captures photoplethysmography (green LED, 200 Hz). Used for HAS score, HRV, and daytime autonomic profiling.
Polar H10 — ECG: chest strap; gold-standard RR interval measurement (1 kHz). Used for precise RMSSD computation during biofeedback sessions and sleep.

All biometric data is processed on-device and uploaded to Firebase Firestore under users/{uid}/daily_summary and users/{uid}/sleep_nights.

עברית

OGEN משלבת שתי מודליות חיישנים עיקריות:

Verity Sense (Polar) — PPG: נלבש על הזרוע העליונה; לוכד פוטופלתיסמוגרפיה (LED ירוק, 200 הרץ). משמש לציון HAS, HRV ופרופיל אוטונומי יומי.
Polar H10 — ECG: רצועת חזה; מדידת מרווחי RR בסטנדרט זהב (1 קילוהרץ). משמש לחישוב RMSSD מדויק במהלך סשנים של ביופידבק ושינה.

כל הנתונים הביומטריים מעובדים במכשיר ומועלים ל-Firebase Firestore תחת users/{uid}/daily_summary ו-users/{uid}/sleep_nights.

Metric Definitions

HAS Score autonomic

ציון HAS

au, 0–100

Hyperarousal Score — composite index integrating Heart Rate Reserve (Karvonen 1957), HRV suppression Z-score, and HR trend component. Computed on-device per reading (app layer, not firmware) and averaged per day. Higher values reflect greater sympathetic activation. No normative population data is published for PTSD specifically.

Formula: HAS_i = clamp(HRR_contrib + HRV_contrib + Trend_contrib, 0, 100). HRR via Karvonen (1957): HRR% = (HR − HR_rest) / (HR_max − HR_rest) × 100. HR_max via Tanaka (2001): 208 − 0.7 × age. Source: daily_summary.has_mean.

Resting HRV (RMSSD) HRV

HRV במנוחה (RMSSD)

Root Mean Square of Successive Differences between RR intervals, measured during periods of minimal motion. RMSSD is the most widely validated time-domain HRV index for vagal tone. Higher values reflect greater parasympathetic (vagal) activity. Normal adult resting range: 20–80 ms; PTSD populations often show suppression [1].

Formula: RMSSD = √(mean((RR_{i+1} − RR_i)²)). Computed from Polar H10 ECG RR intervals (preferred) or Verity Sense PPG-derived RR. Source: daily_summary.rest_rmssd. Note: stationary filtering (isStationary flag) is planned for a future build; current values may include motion artifact.

Resting HR HRV

דופק במנוחה

bpm

Mean instantaneous heart rate during stationary monitoring windows. Elevated resting HR (>80 bpm) at rest is associated with higher sympathetic load and impaired recovery [2].

Formula: mean(HR_i | isStationary=true). Source: daily_summary.rest_hr.

Dangerous Calm Events events

אירועי רגיעה מסוכנת

count/day

Episodes where physiological arousal collapses to near-baseline levels despite contextual cues suggesting threat. Reflects dorsal vagal shutdown — the dissociative "freeze" state described in Polyvagal Theory [3]. Elevated rates (>2/day) may indicate dissociative avoidance patterns warranting clinical review.

Formula: count(events.type = 'dangerous_calm'). Source: events subcollection → daily_summary.dangerous_calm_events.

Total Sleep Time (TST) sleep

זמן שינה כולל

minutes

Total duration of sleep epochs in the overnight recording window. Recommended adult sleep: 420–540 minutes (7–9 hours) [4]. PTSD is frequently associated with TST <360 min due to nightmares, hyperarousal-driven awakenings, and early-morning insomnia [5].

Formula: TST = sleepEpochs × 30s. 3-pass HMM staging (ENMO + HR + actigraphy). Source: sleep_nights → daily_summary.tst_minutes.

Sleep Efficiency (SE) sleep

יעילות שינה

% (0–100)

Ratio of TST to time in bed (TIB). SE ≥85% indicates good sleep consolidation; <75% is the clinical insomnia threshold used in CBT-I trials [6].

Formula: SE = sleepCount / total. Stored as 0–1 fraction in Firestore (daily_summary.sleep_efficiency), displayed as %.

WASO sleep

WASO

minutes

Wake After Sleep Onset — total minutes spent awake after first falling asleep. Reflects sleep maintenance difficulty rather than sleep initiation. PTSD elevates WASO through nightmare-induced arousals and nocturnal hyperarousal [5]. Normal: <20 min. Clinical significance: >40 min.

Source: daily_summary.waso_minutes.

Sleep Fragmentation sleep

פיצול שינה

events/hour

Number of sleep-to-wake transitions per hour of sleep. Fragmented sleep is highly prevalent in PTSD, with rates of 10–20 events/hr reported versus <5/hr in healthy controls [7].

Formula: awakenings / TST_hours (kConsolidateGap=5 epochs)(standard actigraphy consolidation; Sadeh 1994). Source: daily_summary.fragmentation_per_hour.

Nocturnal HR & HRV sleep

דופק לילי ו-HRV לילי

bpm / ms

Mean heart rate and RMSSD across all sleep epochs. Elevated nocturnal HR and suppressed nocturnal HRV are robust biomarkers of ANS dysregulation in PTSD [8]. Normal nocturnal RMSSD: 30–100 ms; PTSD often <30 ms.

Sources: daily_summary.night_mean_hr, daily_summary.night_mean_rmssd.

Statistical Methods

English

All group comparisons are labeled "exploratory analysis". The following methods are implemented in stats.js and validated against scipy reference fixtures.

Descriptive Statistics

Median, mean, SD, IQR, Q1, Q3. Bootstrap 95% CI of median: 2000 resamples with mulberry32 PRNG (seed=0) for exact reproducibility across JS and Python.

Mann-Whitney U

Non-parametric two-sample test. Normal approximation with continuity correction (use_continuity=True), two-sided. Matches scipy.stats.mannwhitneyu.

Effect Size

Cliff's delta: proportion of (x>y) minus (x<y) comparisons. Range [−1, +1]; |d|>0.47 = large.
Hedges' g: pooled-SD standardized mean difference with correction factor 1 − 3/(4(n_x+n_y−2)−1).

Theil-Sen Slope

Robust monotone trend estimator. Slope = median of all pairwise slopes; intercept = median(y) − slope × median(x). Matches scipy.stats.theilslopes.

Spearman Correlation

Rank-based correlation with t-distribution p-value (t = r√((n−2)/(1−r²))). Matches scipy.stats.spearmanr.

Benjamini-Hochberg FDR

Controls False Discovery Rate when multiple metrics are compared simultaneously. Backward cumulative-minimum pass over sorted p-values. Matches scipy.stats.false_discovery_control(method='bh').

Robust Z-Score

MAD-based z-score: z = (x − median(anchor)) / (1.4826 × MAD). Resistant to outliers; used for within-patient progress tracking against personal anchor.

עברית

כל השוואות הקבוצות מסומנות "ניתוח אקספלורטורי". השיטות הבאות מיושמות ב-stats.js ומאומתות מול נתוני ייחוס scipy.

סטטיסטיקה תיאורית

חציון, ממוצע, SD, IQR, Q1, Q3. רווח בטחון 95% לחציון בבוטסטראפ: 2000 מדגמים חוזרים עם PRNG mulberry32 (seed=0).

Mann-Whitney U

מבחן שתי מדגמים לא פרמטרי. קירוב נורמלי עם תיקון רציפות, דו-צדדי.

גודל אפקט

דלתא של קליף: יחס השוואות (x>y) פחות (x<y). טווח [−1, +1]; |d|>0.47 = גדול.
g של הדג'ס: הפרש ממוצעים מנורמל עם SD מאוחד ומקדם תיקון.

BH FDR

שליטה בשיעור גילויים שגויים בהשוואות מרובות. מעבר מינימום מצטבר לאחור על p-values ממוינים.

Z-ציון חזק

z-ציון מבוסס MAD: z = (x − חציון(עוגן)) / (1.4826 × MAD). עמיד בפני ערכים חריגים.

Algorithm Notes

Sleep Staging (on-device)

3-pass algorithm: (1) per-epoch ENMO ≤ 40 mg + HR < daytimeBaseline × 0.92 + rolling HR SD ≤ 7 BPM over 8 epochs; (2) HMM Viterbi (P(stay_sleep)=0.97); (3) Cole-Kripke actigraphy smoothing. Epoch length: 30 seconds.

Multi-pass wearable actigraphy algorithm using accelerometry-derived activity levels, heart rate criteria relative to daytime baseline, a Hidden Markov Model (HMM) for sleep-state sequence coherence, and Cole-Kripke actigraphy smoothing. Epoch length: 30 seconds. Exact algorithm parameters are proprietary — detailed methodology available to OGEN team upon partnership review.

Dangerous Calm Detection

Fires when all 9 conditions are simultaneously true: ACC fresh <5s, ENMO below rest ≥180s, no cadence, HR > baseline+20 BPM, hrv_z_rest < −2.5, ActivityState=REST, HRV_VALID, RMSSD plausible, not in post-exercise 10-min window. Polar ECG only.

Multi-condition autonomic freeze detector combining accelerometry-confirmed sustained immobility, heart rate elevated above personal resting baseline, HRV substantially below personal resting baseline, absence of movement cadence, and valid cardiac data quality. All conditions must be simultaneously true throughout the detection window. Exact thresholds are proprietary — detailed methodology available to OGEN team upon partnership review. Polar ECG only.

Personal Baseline & Anchor

The personal baseline is computed from the first 14 complete (non-incomplete) nights starting from the anchor date. This anchor can be set per-patient by the clinician. The "Anchor Δ" column shows percent change from anchor-period median to current-period median — a within-patient effect size that is robust to inter-patient variability.

Sleep Regularity

Planned metric: standard deviation of sleep midpoint across nights (SRI index). More regular sleep timing is associated with better autonomic recovery and reduced PTSD symptom severity.

References

All citations are verified by DOI or PubMed.

[1] Shaffer F & Ginsberg JP (2017). An Overview of Heart Rate Variability Metrics and Norms. Front Public Health, 5, 258.
doi:10.3389/fpubh.2017.00258
[2] Buckley U & Shivkumar K (2016). Stress-induced cardiac arrhythmias. Trends Cardiovasc Med, 26(1), 78–80.
doi:10.1016/j.tcm.2015.05.001
[3] Porges SW (2011). The Polyvagal Theory. W. W. Norton & Company.
ISBN 978-0-393-70700-7
[4] Watson NF et al (2015). Recommended Amount of Sleep for a Healthy Adult. Sleep, 38(6), 843–844.
doi:10.5665/sleep.4716
[5] Spoormaker VI & Montgomery P (2008). Disturbed sleep in post-traumatic stress disorder. Sleep Med Rev, 12(3), 169–184.
doi:10.1016/j.smrv.2007.08.008
[6] Morin CM et al (2006). Psychological and behavioral treatment of insomnia. Sleep, 29(11), 1398–1414.
doi:10.1093/sleep/29.11.1398
[7] Maher MJ et al (2006). Sleep disturbances in patients with post-traumatic stress disorder. CNS Drugs, 20(7), 567–590.
doi:10.2165/00023210-200620070-00003
[8] Cohen H et al (1997). Power spectral analysis of heart rate variability in PTSD patients. Biol Psychiatry, 41(5), 503–507.
doi:10.1016/s0006-3223(96)00525-2
[9] Lanius RA et al (2010). Emotion modulation in PTSD: Clinical and neurobiological evidence for a dissociative subtype. Am J Psychiatry, 167(6), 640–647.
doi:10.1176/appi.ajp.2009.09081168

OGEN Clinical Analytics · Research Edition ·

OGEN · Research Dashboard

Descriptives by Group

Box / Violin Plot

Forest Plot — Effect Sizes (Cliff's delta, BH-adjusted)

Cohort Trajectory — —

Per-Patient Theil-Sen Slopes

Responder Analysis

Spearman Correlation Matrix

Data Quality Summary

Coverage Distribution

Data Freshness

HAS Zone Distribution

Rest HR vs HRV

HRV vs HAS Scatter

Baseline Drift

Event Rates per 1,000 Monitored Hours

Crisis PPV (where feedback exists)

Event Timeline

Reserve vs Energy Check-in

Sleep Regularity (SD of Midpoint)

OGEN · Clinical Science

Physiological Signals

Metric Definitions

Statistical Methods

Descriptive Statistics

Mann-Whitney U

Effect Size

Theil-Sen Slope

Spearman Correlation

Benjamini-Hochberg FDR

Robust Z-Score

סטטיסטיקה תיאורית

Mann-Whitney U

גודל אפקט

BH FDR

Z-ציון חזק

Algorithm Notes

Sleep Staging (on-device)

Dangerous Calm Detection

Personal Baseline & Anchor

Sleep Regularity

References

OGEN Analytics

Reset Password

Clinical Alerts

Validation Dataset — 23 Synthetic Testers

Zone Distribution — 23 Testers

Trend Distribution

HRV vs Autonomic Load (HAS %)

Cohort Analysis

Clinical Reference

Avg HRV

Avg Stress

Crisis Readings

Baseline Drift

REM HR Δ

Status Badge

Stress Index Trend (daily avg)

Zone Distribution

Avg HRV Trend (ms)

PHYSIOLOGICAL_STATUS Distribution (Sam sessions)

Daily Detail

Observation Notes (auto-generated · manual review only · no auto-apply)

Clinical Digest — Caseload Triage View

—

Personal Baseline Metrics

Trends

Daily HAS Mean

Rest RMSSD

Total Sleep Time (min)

Sleep Fragmentation/hr

Sleep Fragmentation (7 days)

Sleep Nights

Statistical Summary (all data since anchor)

Autonomic Profile

Events (last 30 days)

Clinical Notes

Add Note

Clinical Focus Metrics (max 5)

AI Clinical Insight