Skip to main content
Fig. 5 | Journal of NeuroEngineering and Rehabilitation

Fig. 5

From: Semi-automated approaches to optimize deep brain stimulation parameters in Parkinson’s disease

Fig. 5

Efficiency comparison analysis of different methods. Frequency range (bottom) after each iteration of Bayesian optimization (blue circles) and two post-hoc frequency selection methods. The first post-hoc frequency selection method sampled approximately equal intervals 5Hz (green squares). The second method randomly selected frequencies (orange triangles), and the shaded light orange region indicates the standard deviation of 50 simulations at each iteration. dashed line indicates the frequency range of the full brute-force GP model. The frequency range measured the distance between the optimal frequency and the lowest frequency that is contained within the optimal frequency’s mean and standard deviation

Back to article page