Neuronal Ensemble Decoding Using a Dynamical Maximum Entropy Model
Figure 5
A comparison of decoding error and uncertainty under various degrees of stimulus direction difference between the proposed (new) and the traditional maximum entropy decoders. (a) Both decoders demonstrate a similar trend of higher error rates for greater input difference. Except for the condition, the proposed method yields significantly lower error rates (-test, ). (b) A sensible increase of uncertainty is shown for both methods as the input difference grows. The proposed method contains greater uncertainty (-test, ) for all cases.
