Background: Recent technological advances in deep brain stimulation (DBS) (e.g., directional leads, multiple independent current sources) lead to increasing DBS-optimization burden. Techniques to streamline and facilitate programming could leverage these innovations. Objective: We evaluated clinical effectiveness of algorithm-guided DBS-programming based on wearable-sensor-feedback compared to standard-of-care DBS-settings in a prospective, randomized, crossover, double-blind study in two German DBS centers. Methods: For 23 Parkinson’s disease patients with clinically effective DBS, new algorithm-guided DBS-settings were determined and compared to previously established standard-of-care DBS-settings using UPDRS-III and motion-sensor-assessment. Clinical and imaging data with lead-localizations were analyzed to evaluate characteristics of algorithm-derived programming compared to standard-of-care. Six different versions of the algorithm were evaluated during the study and 10 subjects programmed with uniform algorithm-version were analyzed as a subgroup. Results: Algorithm-guided and standard-of-care DBS-settings effectively reduced motor symptoms compared to off-stimulation-state. UPDRS-III scores were reduced significantly more with standard-of-care settings as compared to algorithm-guided programming with heterogenous algorithm versions in the entire cohort. A subgroup with the latest algorithm version showed no significant differences in UPDRS-III achieved by the two programming-methods. Comparing active contacts in standard-of-care and algorithm-guided DBS-settings, contacts in the latter had larger location variability and were farther away from a literature-based optimal stimulation target. Conclusion: Algorithm-guided programming may be a reasonable approach to replace monopolar review, enable less trained health-professionals to achieve satisfactory DBS-programming results, or potentially reduce time needed for programming. Larger studies and further improvements of algorithm-guided programming are needed to confirm these results.

Understanding the way stimulus properties are encoded in the nerve cell responses of sensory organs is one of the fundamental scientific questions in neurosciences. Different neuronal coding hypotheses can be compared by use of an inverse procedure called stimulus reconstruction. Here, based on different attributes of experimentally recorded neuronal responses, the values of certain stimulus properties are estimated by statistical classification methods. Comparison of stimulus reconstruction results then allows to draw conclusions about relative importance of covariate features. Since many stimulus properties have a natural order and can therefore be considered as ordinal, we introduce a bivariate ordinal probit model to obtain classifications for the combination of light intensity and velocity of a visual dot pattern based on different covariates extracted from recorded spike trains. For parameter estimation, we develop a Bayesian Gibbs sampler and incorporate penalized splines to model nonlinear effects. We compare the classification performance of different individual cell covariates and simple features of groups of neurons and find that the combination of at least two covariates increases the classification performance significantly. Furthermore, we obtain a non-linear effect for the first spike latency. The model is compared to a naïve Bayesian stimulus estimation method where it yields comparable misclassification rates for the given dataset. Hence, the bivariate ordinal probit model is shown to be a helpful tool for stimulus reconstruction particularly thanks to its flexibility with respect to the number of covariates as well as their scale and effect type.

In a natural environment, detection and estimation of object motion features are crucial for a correct behavior. The brain relies on the activity of retinal ganglion cells (RGC) as the only source of visual information to estimate moving stimuli. We analysed simultaneously recorded responses of RGC of isolated turtle retinas to a moving stimulus. The stimulus consisted of a pattern of squares that moved with a constant velocity for 500 ms and then changed abruptly to one of nine possible velocities in a pseudo-random fashion as described in [1]. Spike-cost based metrics [2] were applied to the responses of single RGC to test the relevance of spike rate and spike timing precision in the encoding of velocity. Furthermore, an extension of the method that considers the simultaneous activity of several neurons Our results show that responses of single RGC allow for discrimination of different velocities based on the spike count rather than on the spike timing precision. In general, higher velocities were better discriminated than lower ones. Classification performance of certain velocities depends on the cell type. Responses of non-direction selective cells (NDSC) allow for good classification of all speeds but fail in the discrimination of movement direction (Figure 1a), whereas activity of direction selective cells (DSC) allow for better classification of velocities in the preferred direction but worse for velocities in the antipreferred direction (Figure 1b). For all types of cells, mean velocity classification performance reached ~35%. When the activity of different RGC is combined, classification performance of velocity improves with the number of RGC and only if the neuron of origin of each spike is known (Figure 1c). Nevertheless, mean classification performance is lower than the one expected by summing the performances of single cells, suggesting some redundancy in the encoding of velocity parameters also by different types of RGC.

No abstract