Skip to main content
. 2009 Apr 29;29(17):5573–5586. doi: 10.1523/JNEUROSCI.4438-08.2009

Table 4.

Polynomial fits to model properties

Property RMSE for constant fita Number of coefficientsb Test RMSEc Test R2d
Input conductance (nS) 8.69 112 2.88 0.89
Resting membrane potential (mV) 2.12 107 0.48 0.94
Resting spike rate (Hz) 3.44 135 0.82 0.95
Phase of burst onset (%) 3.20 158 1.85 0.70
Phase of burst offset (%) 2.64 30 1.95 0.43
Spike rate in burst (spikes/cycle) 3.40 140 0.92 0.93
Slow-wave amplitude (mV) 3.54 151 0.65 0.96
Peak slow-wave potential (mV) 2.51 122 0.54 0.95

aThe root mean-squared error (RMSE) for the constant that best fits the model property. This is also equal to the square root of the variance of that property, calculated over all 1304 LP models.

bThe number of polynomial terms used in the best fit, as determined using cross-validation (see Materials and Methods).

cThe RMSE for the final polynomial fit, as evaluated on the test set.

dThe, R2 for the final polynomial fit, as evaluated on the test set.