Table 2.
Data collection methods, their Pros and Cons.
| References | Modality | Pros | Cons |
|---|---|---|---|
| Rosenberg et al. (6) | Wearable ECG | Detect stress with 90% accuracy | Less effective during pain and non-stationary situations |
| Blood et al. (7) | Holter ECG | Effectively detect depression and HRV | Accuracy of results |
| Molina et al. (8) | 12-lead ECG | Accurate correlation between HRR and HRV | May cause scar |
| Leti and Bricout (9) | Polar RS 800 | Detect fatigue and HRV in motion | Accuracy of Results |
| Walker et al. (10) | GE Light ECG | Effectively analyze Noise exposure and HRV | Did not detect correlation between noise and BP |
| Wang et al. (11) | Wearable ECG | Discriminate between CHF and NSR with 91.3% acc | RMSSD is not accurate |
| Huang et al. (12) | 12-lead ECG | Effectively determine HRV due to stroke and hemodialysis | LF/HF ratio is not accurate |
| Pinheiro et al. (13) | PTB recorder | Determine prognosis of patients following MI | Cannot deduce causality behind results |
| Toni et al. (14) | Clickholter ECG | Detect HRV in motion due to antidepressants and exercise | LF/HF, RR are not accurate |
| Shi et al. (15) | RM6240B ECG | Effectively discriminate between HRV of happiness and sadness | RMSSD, pNN50 and SampEn are not accurate |
| Howells et al. (17) | MP150 Biopac | Accurately analyzed HRV due to meditation and BD wirelessly | Results lacked most ECG measures |
| Rios et al. (18) | Gear S, PPG | Possibly recognize drowsiness while in motion | No results were obtained |
| Jung et al. (19) | ECG sensor | Wireless analysis of HRV due to drowsiness and fatigue | Accuracy of results |
| Georgiou et al. (21) | ECG,PPG | Analyze HRV with 91-99 % accuracy | Accuracy reduces during motion |
| Gontier (22) | eMotion Faros | Efficiently detect correlation between awareness and HR | Did not find robust correlations |
| Vicente et al. (23) | eXim Pro | Detect drowsiness while in motion | Detect drowsiness with 62% sensitivity |
| He et al. (24) | custom ECG | Detect stress using ulta-short epoch | Accuracy of classification was not revealed |
| Schmidt et al. (25) | RespiBAN Empatica E4 | Detect stress with 93% accuracy | May have resulted from overfitting |
| Cho et al. (26) | Biopac PPG EDA,UIM | Detect stress with 95% accuracy | Not a viable solution in real-life |