Meltdown Minder Knowledge Base

n=909. Population: ASD. Setting: inpatient. Setting: clinical.

Clinical setting. Laboratory. During sleep. Free-living. Real-world.

ABC, VAS, Clinician-rated, Direct observation, Self-report, Event labels.

n=909.

Laboratory.

Event labels.

Laboratory. During sleep.

Event labels.

n=65. Population: ASD. Setting: clinical.

Clinical setting. Laboratory. During sleep. Real-world.

Event labels.

n=65.

Laboratory.

Event labels.

n=62.

Laboratory.

Event labels.

n=62.

n=142.

Laboratory. Real-world.

Self-report, Event labels.

n=142.

n=32.

Laboratory.

Self-report, Event labels.

n=32.

Laboratory.

Event labels.

n=20. Population: ASD. Setting: inpatient.

Laboratory.

VAS, Direct observation, Event labels.

n=20.

n=70. Population: ASD. Setting: inpatient.

Laboratory.

Event labels.

n=70.

n=30.

Laboratory.

Self-report, Event labels.

n=30.

Population: ASD.

Laboratory. Real-world.

Self-report, Event labels.

n=3. Population: ASD.

Laboratory. Real-world.

Event labels.

n=3.

n=25.

Laboratory.

Event labels.

n=25.

n=212.

Laboratory. Free-living. Real-world.

Self-report, Event labels.

n=212.

n=15.

Laboratory.

Event labels.

n=15.

n=335. Population: ASD. Setting: clinical.

Clinical setting. Laboratory. During sleep.

Event labels.

n=335.

Systematic review — no original subject recruitment. Literature search across academic databases.

N/A — review article. Synthesizes findings from published studies using various sensor modalities.

N/A — review article. Draws from studies across clinical, home, lab, and naturalistic settings.

N/A — review article. Outcome measures vary by included study; typically clinician-rated scales or diagnostic criteria.

N/A — review article. Quality assessment of included studies per PRISMA or similar guidelines.

N/A — review article. No primary data collection.

n=70. Age range: 6-17 yr. Population: ASD. Setting: inpatient.

EDA, ACC, TEMP. Wearable/wrist-worn device.

Naturalistic setting. Inpatient unit. Laboratory.

VAS, Direct observation, Event labels.

n=70.

n=70. Age range: 5-19 yr. Population: ASD. Setting: inpatient.

Empatica E4. Wearable/wrist-worn device.

Inpatient unit. Laboratory.

Event labels.

n=70.

n=3. Age range: 6-9 yr. Population: ASD.

EDA, ACC. Wearable/wrist-worn device.

Naturalistic setting. Laboratory. Real-world.

Event labels.

n=3.

n=70. Youth with ASD (ages 6-20) from inpatient psychiatric units. Subset of AIC (Autism Inpatient Collection).

Empatica E4 wristband. EDA (4Hz), PPG (64Hz), ACC (32Hz), TEMP (4Hz). Continuous monitoring for up to 14 days.

Inpatient psychiatric unit. Naturalistic observation during daily activities. No structured tasks. Clinical staff logged aggression events via OAS-M.

OAS-M (Overt Aggression Scale — Modified). Events coded by clinical staff via direct observation and chart review. Inter-rater reliability: κ > 0.80.

Temporal point process modeling. Event times validated against clinical logs. EDA signal quality assessed via standard deviation thresholding.

n=70 enrolled. 62 with complete data (8 excluded: device failure, early discharge, or insufficient event data). Missingness correlated with high-agitation periods.

Missing not at random

Systematic review — no original subject recruitment. Literature search across academic databases.

N/A — review article. Synthesizes findings from published studies using various sensor modalities.

N/A — review article. Draws from studies across clinical, home, lab, and naturalistic settings.

N/A — review article. Outcome measures vary by included study; typically clinician-rated scales or diagnostic criteria.

N/A — review article. Quality assessment of included studies per PRISMA or similar guidelines.

N/A — review article. No primary data collection.

n=15. Healthy adults (mean age 27.5). 3 conditions: baseline, amusement, stress. Recruited from university community.

Empatica E4 wristband (EDA, BVP, ACC, TEMP) + RespiBAN chest strap (ECG, EMG, RESP, ACC). Both devices worn simultaneously.

Laboratory setting. Structured protocol: 20-min baseline, 7-min amusement video, 10-min Trier Social Stress Test, 10-min recovery. Each session ~2 hours.

Self-report (SAM, PANAS, SSSQ). Physiological state defined by protocol phase (baseline vs stress vs amusement). No clinician-rated labels.

Manual artifact inspection. EDA artifacts removed via visual inspection. ECG: Pan-Tompkins for R-peak detection. Accelerometer non-wear detected via zero-variance.

n=15 enrolled. All 15 completed full protocol. No missing data reported. Complete data for all participants across all sensors.

n=1,021. Pooled from 13 public datasets (WESAD, DEAP, DAPPER, etc.). Diverse demographics. No original recruitment.

PPG (from Empatica E4, Polar H7, and other wearables). Sampling rates vary by source dataset (64-256Hz).

Pooled analysis of 13 existing datasets. Original acquisition contexts vary: lab, controlled, and naturalistic settings.

Self-reported emotion labels (valence, arousal, dominance per SAM). Labels normalized across datasets via harmonization procedure.

Signal quality filtering: PPG signals with SNR < 3dB excluded. Inter-beat interval outliers removed via MAD-based thresholding. Harmonization pipeline validated on held-out dataset.

n=1,021 pooled from 13 datasets. Each dataset had independent exclusion criteria. Full details in source dataset publications.

Missing

n=45. Children with ASD (ages 5-12). Recruited from outpatient clinics and community organizations. Parent-reported diagnosis confirmed via ADOS.

Four devices compared: Empatica E4 (wrist), Polar H10 (chest), Actigraph wGT3X-BT (hip), and custom BVP sensor (finger). Simultaneous wear for 24h.

Home environment. Free-living conditions. Parents instructed to maintain normal daily routines. Devices worn for 24 consecutive hours.

Parent-reported comfort/skin irritation logs. Device wear-time validated via accelerometer non-wear detection. No clinical outcome events.

Cross-device comparison of signal quality metrics. Wear-time validation via multiple sensors. Skin irritation assessed via parent diary.

n=45 enrolled. 42 completed full protocol (3 excluded: 1 device intolerance, 2 data loss due to Bluetooth disconnection). 93% complete wear-time data.

Missing

Systematic review — no original subject recruitment. Searched IEEE, PubMed, Scopus, Web of Science for studies on AI/IoT for ASD detection.

N/A — systematic review. Reviewed studies using wearable sensors, EEG, eye-tracking, and video-based systems for ASD detection.

N/A — systematic review. Pulls from published studies across clinical, home, and lab settings.

N/A — systematic review. Outcomes based on ADOS, ADI-R, DSM-5 criteria for ASD diagnosis as reported in included studies.

PRISMA guidelines followed. Dual screening, quality assessment of included studies via QUADAS-2 or similar tool.

N/A — systematic review. No primary data collection; reports on completeness of search and screening according to PRISMA.

N/A — review article. No human subjects. Covers MXene-based biosensor platforms for biomedical diagnostics.

N/A — review article. Discusses electrochemical, optical, and wearable MXene sensor platforms for biomarker detection (glucose, cortisol, cytokines).

N/A — review article. Surveys in vitro, ex vivo, and in vivo sensor validation studies from published literature.

N/A — review article. No outcome events. Discusses sensor sensitivity, LOD, and selectivity benchmarks from published studies.

N/A — review article. Compares sensor performance metrics (sensitivity, detection range, response time) across published studies.

N/A — review article. No primary data collection.

n=70. Youth with ASD (ages 6-20) from inpatient psychiatric units. 75% male, 85% minimally verbal. Recruited from 4 sites as part of AIC (Autism Inpatient Collection).

Empatica E4 wristband. Sensors: EDA (4Hz), PPG (64Hz), ACC (32Hz), TEMP (4Hz). Data streamed via Bluetooth to study tablet.

Inpatient psychiatric unit. Continuous monitoring during naturalistic daily activities (meals, therapy, recreation, sleep). No structured tasks. Duration: up to 14 days per participant.

Aggression events defined per OAS-M (Overt Aggression Scale — Modified). Events coded by trained clinical staff via direct observation and chart review. Inter-rater reliability: κ > 0.80.

Manual artifact rejection by trained research assistants. EDA artifacts identified via sudden drops > 5μS. ACC non-wear detected via zero-variance windows. Cross-modal validation between EDA and HRV peaks.

n=70 enrolled. 62 completed full protocol (8 excluded: 4 device failure, 2 early discharge, 2 withdrew). Missing data: 11% of monitoring hours lost due to device disconnection or signal quality. Missingness correlated with high-agitation periods (MNAR).

Missing not at random

Population: ASD.

Laboratory.

Event labels.

Setting: clinical.

Clinical setting. Laboratory.

Event labels.

Laboratory.

Event labels.

n=1.

Laboratory.

Event labels.

n=1.

Setting: clinical.

Clinical setting. Laboratory.

Event labels.

Population: ASD.

Laboratory. Real-world.

Event labels.

Laboratory.

Event labels.

Laboratory.

Event labels.

Laboratory.

Event labels.