Personalized medicine in deep brain stimulation through utilization of neural oscillations.

Neurology 2012;78:1900–1901 The neural underpinnings for normal and abnormal basal ganglia functions and functional connectivities remain unknown, although several important pieces to the puzzle have emerged over 3 decades. First was the identification of segregated basal ganglia circuits.1 This discovery was closely paralleled by the realization that these circuits communicated through the use of a special physiologic language,2 which proved to be more complex than initially thought, as it was not the simple firing rate of neurons, but rather specific patterns of oscillatory neuronal discharges that were important.3 The most recent realization is the oscillation model, according to which the oscillatory neuronal discharges in specific frequency bands dictate specific motor behaviors.4 In Parkinson disease, increased endogenous frequencies in the (4 –10 Hz) and (11–30 Hz) bands recorded from the subthalamic nucleus (STN) region are associated with worsening of motor symptoms. These bands have been referred to as antikinetic or bad frequency bands. In contrast, frequencies (31–100 Hz) are associated with motor improvements, and are referred to as prokinetic, or good frequency bands.5 Complicating the picture, the peak frequencies in each of the 3 bands recorded from the STN region may vary widely among persons.6 Tsang and colleagues,7 in this issue of Neurology, ask whether these specific oscillatory frequencies could be utilized to tailor a personalized approach to STN deep brain stimulation (DBS). The current clinical practice of DBS is empirical, and utilizes a high-frequency 100 Hz signal for therapeutic stimulation of the STN region.8 According to the oscillation model, therapeutic STN DBS in the (31–100 Hz) frequency band should artificially drive a prokinetic circuit. Stimulation in (4–10 Hz) and (11–30 Hz) bands should worsen the motor response. These stimulation effects could possibly be more robust if STN DBS was delivered at specific frequencies that were individualized for specific patients and specific symptoms. In this proof of concept study, individually defined medicationdependent and movement-related peak frequencies across , , and bands of stimulation were determined and applied, and correlated with potential related clinical motor responses. Individual frequencies were ascertained directly from the DBS leads following insertion into STN, sampled within the first month after surgery before connection to the final chest-based battery source. For each subject, the peak frequency was determined that showed a reduction in the antikinetic band ( and frequencies) and an increase in the prokinetic band ( frequency) during medication “on” recordings, as compared to medication “off ” periods. Similarly, peak frequencies that reduced the antikinetic and increased in the prokinetic band for movement periods as compared to premovement periods were recorded when subjects performed self-initiated and externally triggered wrist movements. These frequencies in the , , and bands were called individualized frequencies. The authors cleverly employed DBS at these “individually” determined frequencies, and compared their results to the empirically chosen high-frequency stimulation that was utilized by the treating team. This testing was performed at a minimum of 3 months after the surgery in order to avoid postsurgical effects (implantation and microlesion effects). The investigators concluded that the motor benefits of DBS at individual frequencies were comparable to those obtained with high-frequency stimulation. These findings were puzzling, as one would have hypothesized that stimulation at individualized prokinetic rhythms would have been superior to conventional high-frequency stimulation. The empirically determined high frequencies that were used for chronic stimulation did not reflect as harmonics of the prokinetic frequencies, and therefore the authors suggested an intriguing possibility that the basal ganglia circuit could have more than one prokinetic frequency band. If this suggestion proves true, it will have the potential to alter future therapeutic approaches.