Table 3. Individual and mean regression slopes for ECoG measures derived from the univariate linear regression models.
The one-sample t-test revealed that the mean regression slope for very high-frequency gamma (100–200 Hz), high-frequency gamma (64–100 Hz), low-frequency gamma (32–64 Hz), beta (16–32 Hz), and sigma spectral amplitudes (12–16 Hz) were significantly greater than zero.
ECoG Measure | Regression Slope (R2) for Patient 1 | Regression Slope (R2) for Patient 2 | Regression Slope (R2) for Patient 3 | Regression Slope (R2) for Patient 4 | Regression Slope (R2) for Patient 5 | Regression Slope (R2) for Patient 6 | Mean Regression Slope (Mean R2 among Six Children | 95%CI for the Mean Regression Slope |
---|---|---|---|---|---|---|---|---|
'Delta amplitude'; β D | 0.00008 (0.0002) |
0.0013*** (0.14) |
0.00035* (0.046) |
0.00044* (0.046) |
0.00071*** (0.096) |
−0.00087*** (0.14) |
0.00033 (0.078) | −0.0004, 0.0011 |
'Theta amplitude'; β T | 0.0013 (0.025) |
0.0018** (0.078) |
0.0015*** (0.17) |
0.0015** (0.061) |
0.00056 (0.030) |
−0.0015*** (0.14) |
0.00088 (0.084) |
−0.0004, 0.0022 |
'Alpha amplitude'; β A | 0.0088** (0.13) |
0.0029** (0.048) |
0.0024*** (0.23) |
0.0029* (0.045) |
0.0012*** (0.10) |
−0.0009 (0.030) |
0.0029 (0.097) |
−0.0005, 0.0063 |
'Sigma amplitude'; β S | 0.0161*** (0.16) |
0.0087*** (0.10) |
0.0035*** (0.22) |
0.0041 (0.023) |
0.0026*** (0.22) |
0.0013* (0.041) |
0.0060* (0.13) |
0.0002, 0.012 |
'Beta amplitude'; β B | 0.0074*** (0.22) |
0.0026* (0.047) |
0.0018*** (0.21) |
0.0025* (0.034) |
0.0009*** (0.11) |
0.0017*** (0.25) |
0.0028* (0.15) |
0.0004, 0.0052 |
'Low-frequency gamma amplitude'; β LG | 0.0049** (0.11) |
0.0036** (0.061) |
0.0037*** (0.17) |
0.0037** (0.074) |
0.0011*** (0.086) |
0.0020*** (0.28) |
0.0032** (0.13) |
0.0017, 0.0046 |
'High-frequency gamma amplitude'; β HG | 0.0094** (0.12) |
0.012*** (0.23) |
0.0094*** (0.19) |
0.0067*** (0.098) |
0.0036*** (0.14) |
0.0039*** (0.19) |
0.0075** (0.16) |
0.0040, 0.011 |
'Very high-frequency gamma amplitude'; β VHG | 0.018*** (0.15) |
0.018*** (0.30) |
0.0055*** (0.18) |
0.0079*** (0.094) |
0.0032*** (0.26) |
0.0047*** (0.15) |
0.0095* (0.19) |
0.0025, 0.017 |
'Spike frequency'; β SP | −0.081*** (0.37) |
0.010 (0.017) |
0.006*** (0.15) |
−0.018 (0.014) |
0.006*** (0.29) |
−0.010* (0.059) |
−0.015 (0.15) |
−0.050, 0.021 |
p-value<0.05.
p-value<0.01.
p-value<0.001.
Very high-frequency gamma amplitude measure as a predictor had the mean R-square among six patients to be 0.19 in the univariate regression analysis; this observation suggests that about one-fifth of variance of glucose metabolic pattern can be explained by very high-frequency gamma amplitude alone.