mesoscopic region
Last edited: August 8, 2025the mesoscopic region is the regions far away from equilibrium points—which is really hard
This is also why Poincare invented topo.
metabolism
Last edited: August 8, 2025Methods
Last edited: August 8, 2025MFA Disfluency Measurement
Last edited: August 8, 2025Applying the MFA aligner upon the Pitt (cookie only) data and performing statistics upon the calculated disfluency information. The ultimate goal is to replicate Wang 2019.
The code is available here.
The (unvalidated, draft) results are reported below:
Mean value reported, standard deviation in parens. For our data, \(N=422\), cases balanced.
| Variable | AD (Pitt, ours) | MCI (Wang) | Control (ours) | Control (Wang) |
|---|---|---|---|---|
| Silence Duration | 28.10 (21.28) | 13.55 (5.53) | 18.06 (12.52) | 7.71 (5.03) |
| Speech Duration* | 23.77 (14.11) | 46.64 (5.79) | 27.23 (15.3) | 53.63 (7.82) |
| Voice-Silence Ratio | 1.79 (4.88) | 4.43 (2.78) | 5.78 (31.95) | 10.11 (6.05) |
| Verbal Rate | 1.59 (0.61) | 1.56 (0.40) | 1.989 (0.51) | 1.91 (0.43) |
*speech duration would obviously vary with file length
MFA Performance Statistics
Last edited: August 8, 2025- Lanzi WNL (August 12) 1%. Selection Seed 7. Houjun. 82.64% ± 4.48% with a 95% confidence.
- Lanzi MCI (August 12) 1%. Selection Seed 7. Houjun. 78.70% ± 7.85% with a 95% confidence.
- Lanzi WNL (August 13) 1%. Selection Seed 7; syllabic balanced. Houjun. Within which, 90.97%±3.40% of multi-syllabic words were correctly identified 86.28%±4.08% of mono-syllabic words were correctly identified 88.63%±2.65% of all words were correctly identified at a confidence interval of 95% based on a single-variable t test.
- Lanzi MCI (August 13) 1%. Selection Seed 7; syllabic balanced. Houjun. Within which, 76.85%±8.08% of multi-syllabic words were correctly identified 72.22%±8.58% of mono-syllabic words were correctly identified 74.54%±5.86% of all words were correctly identified at a confidence interval of 95% based on a single-variable t test.
Lanzi WNL (August 13) 1%. Selection Seed 7; syllabic balanced; 3-tier labeling. Houjun. Within which, 96.75%±2.10% of multi-syllabic words were correctly identified 90.61%±3.46% of mono-syllabic words were correctly identified 93.68%±2.03% of all words were correctly identified at a confidence interval of 95% based on a single-variable t test.