Intraoperative Experiments

Modulation of Neuronal Ensemble Activity During Movement Planning in Parkinson’s Disease Patients Undergoing Deep Brain Stimulation


Introduction: Multi-unit cortical recordings in non-human primates have demonstrated plan-related neuronal activity in dorsal premotor cortex which is direction-specific.  We hypothesized that similar activity could be recorded from frontal cortex (area 6a) in patients undergoing deep brain stimulation for Parkinson’s disease.

Methods: The experimental setup developed for non-human primates was fully replicated in the operating room. Following opening of the burrhole, a 4x4 mm array of 100 silicon microelectrodes 1.0 - 1.5 mm in length was introduced into the brain using a pneumatic inserter.  Recordings were made during performance of a delayed-reach task in which patients made either leftward or rightward reaches toward previously displayed targets following a “go” cue.  Targets were rear-projected onto a plexiglas screen.  A light-emitting diode was attached to the index finger contralateral to the array implant, and reaches were recorded by tracking the LED with a Polaris camera system. 

Results: Local field potentials recorded from the array showed clear modulation during movement preparation and active movement.  Modulation was most pronounced in the 20-30 Hz band, although other frequencies were also variably modulated throughout the task.

Conclusions: Intracortical recording from human frontal cortex (area 6a) in Parkinson’s patients shows modulation which is similar to that obtained from dorsal premotor cortex of nonhuman primates.  Further characterization of these neuronal signals during the application of deep brain stimulation will lead to better understanding of cortical mechanisms in Parkinson’s disease.



Methods & Materials

Patients undergoing STN DBS for advanced Parkinson’s disease signed an IRB-approved consent form for participation in the study.  Prior to surgery, titanium bone fiducials were placed for a standard frameless DBS approach and 1.25 mm thick CT slices were obtained throughout the entire cranial volume.  These images were fused with both a volumetric contrast-enhanced SPGR MRI and a coronal T2-weighted slab containing the STN.  Targeting was carried out in a standard fashion using the Stealth FrameLink software package (Medtronic, Inc.).  Once in the operating room, a temporary reference was attached non-invasively to the patients head and an entry point planned on the crown of a cortical gyrus approximately 35 mm lateral to the midline in premotor cortex (area 6a).  Following sterile prep and drape, a 14 mm burrhole was made and the NeXframe trajectory guide (Medtronic, Inc.) was attached to the skull.  A custom-made anodized aluminum ring, which served as a mount for array insertion equipment, was attached to the NeXframe (figure 1).

The dura was then opened, and an array of 100 silicon microelectrodes 1.5 mm long and spaced 0.4 mm apart (figure 2) was placed on the cortical surface.  The interface pedestal was attached to the annular ring.  A pneumatic inserter wand was then used to provide a brief, precisely controlled tap to the back of the electrode array, driving it into the cortex (figure 1).  The final appearance of the inserted array is shown in figure 3.  The array was then connected to a data acquisition system (NeuroPort, Cyberkinetics Inc.) which provided on-line amplification, recording, and processing of neural signals. 

The patient was then asked to complete a delayed reaching task. In this task, the patient was asked to touch a center point on a plexiglas screen on which various targets were rear-projected (figure 4). A target was then presented at a distant location. Once a cue to “go” had been given, the subject made a reach to the target. Various configurations of target and delay periods were presented.  The experimental paradigm replicated that used in non-human primates (figure 6).

At the completion of the task and recording, the probe was removed slowly from the brain tissue and the cortex inspected for any damage. The remainder of the DBS operation proceeded as usual.

Neural data was analyzed offline and correlated with reach movements as recorded during surgery. This data was then compared with data obtained from an identical task performed in nonhuman primates as described above.




Successful array insertion was accomplished in 4/6 cases, with two cases being compromised by the inability to find a suitable area of cortex.  In 3 cases, following successful insertion, no recordings could be made due to connection problems (1 case) or insurmountable issues with electrical noise (2 cases).  In the remaining case, we were able to successfully record local field potentials during the behavioral task.

Detailed neurological examination before and after array insertion showed no detectable deficit.  Successful insertions were associated with tiny punctures in the cortex in all cases (related to the electrode insertion sites), with a hemorrhage noted beneath the array in one case.  No complications resulted from array insertion.

Modulation of local field potentials was clearly demonstrated during pre-movement planning in 3 of the 10 trials.  This modulation occurred most strongly in the 20-30 Hz power spectrum (figure 6).  These 3 trials were extracted from the complete spectrogram and are presented in figure 7.  Modulation was also demonstrated during arm movement in 2/10 trials.




Parkinson’s disease has been characterized as a disorder of basal ganglia functioning secondary to depletion of dopamine in the substantia nigra.  These abnormalities have been unequivocally demonstrated by electrophysiological recordings in both animals and humans, and have led to the successful development of deep brain stimulation of the subthalamic nucleus (STN DBS).  However, very little work has focused on the role of the cortex in the generation of Parkinsonian signs and symptoms.  Given the long-held assumption that the motor cortex represents the “final common pathway” for generating movements, investigation of cortical neurophysiology in Parkinson’s disease could provide valuable insights into the relationships between cortical and basal ganglia control of movement.

How does the brain generate rapid and accurate movements of the body? Much work has focused on the role of sensory feedback and internal models in optimizing control signals during movement. However, it is quite possible that the optimization of control signals begins during motor preparation, before movement begins.  Using a delayed reach paradigm in non-human primates, delay-period or ‘preparatory’ activity can be observed in primary motor cortex (M1), and is even more prevalent in dorsal premotor cortex (PMd).  It is often suggested that such activity is related to motor programming.  Preliminary work suggests that such optimization does indeed occur during motor preparation.  


Recordings were made from PMd of two rhesus monkeys.  In the first set of experiments, we asked whether trial-by-trial variability in the state of preparatory activity correlates with trial-by-trial variability in the actual movement.  We showed that a considerable portion of this variability is due to variability in preparatory activity.  In particular, delay-period activity was predictive of the natural variability in the peak speed of the upcoming movement (p<0.001 for both monkeys).  These data indicate that the state of preparatory activity has consequences for the upcoming movement (Churchland & Shenoy 2004; Afshar et al. 2004).

In order to investigate the role of the human cerebral cortex in planning and execution of movements in Parkinson’s patients, we replicated this experimental setup in the operating room during STN DBS procedures for advanced Parkinson’s disease.

Figure 1. Hardware for array insertion.  Note the annular ring attachment with inserter wand and pneumatic cable. The ring attaches to the NeXframe trajectory guidance platform (shown in figure 3).


Figure 2.  Electron micrograph of Cyberkinetics silicon microelectrode array. Scale bar = 2mm.



Figure 3.  The array has been inserted into the cortex.  The annular ring remains attached to the NeXframe (clear plastic tower) but is outside of the field of view. 


Figure 4. A patient positioned for surgery with the plexiglas screen on which targets will be projected, and the two monitors of the Tempo behavioral control system in the background.  A C-arm fluoroscope provides confirmation of DBS lead position.  The reference system with reflective spheres (attached to head) provides localization for the navigational system.




To our knowledge, this is the first demonstration of plan-related activity recorded directly from neuronal ensembles in human premotor cortex. These recordings show modulation which is similar to that obtained from dorsal premotor cortex of nonhuman primates.  Further characterization of these neuronal signals during the application of deep brain stimulation will lead to better understanding of cortical mechanisms in Parkinson’s disease.  Future experiments will aim to correlate symptomatology with cortical ensemble activity and will explore the relationship of cortical ensembles to single-unit activity in the STN and other nuclei of the basal ganglia.

Precisely replicating the experimental setup used for characterizing plan-related activity in non-human primates has several advantages. It allows for immediate application of many years of experience in signal processing and analysis. It encourages the same experimental questions to be asked in both the more controlled laboratory setting and the more clinically-relevant O.R. setting.  The ability to directly translate clinical findings from the primate laboratory to the operating room and back again provides a rich environment for the exploration of the neurophysiology of both normal and abnormal movement.



Figure 5. Spectrogram of LFP of PD patient performing delayed reach task. Vertical lines signify different events during the task. Yellow lines are central touch point presentation; purple and dark green lines are presentation of target (purple is a rightwards target, while dark green is a leftwards target); green lines are go cue; red lines are end of trial.


Figure 6. Spectrogram of a single trial, showing 1.) presentation and acquisition of the central fixation target, 2.) target cue presentation, 3.) GO cue presentation, 4.) end of movement period. 


Figure 7. Three trials exhibiting clear modulation in the 20-30 Hz frequency band (blue ovals) during reach planning (between target presentation at purple line and GO cue at black line).