Mood Response to Deep Brain Stimulation of the Subthalamic Nucleus in Parkinson's Disease

Forty-two individuals with PD and bilateral STN DBS were recruited from the Movement Disorders Center at Washington University in St Louis. Participants met diagnostic criteria for clinically definite PD on the basis of established criteria,^25,26 including benefit from L-dopa; they had no evidence of dementia from clinical history and performance on either the MMSE (≥24/30) or the Mattis Dementia Rating Scale (≥130/148) obtained as part of the pre-surgical neuropsychological evaluation. Participants had no evidence of other serious neurologic diagnoses (e.g., stroke, head injury) by history or examination. DBS therapy was optimized clinically before recruitment in the study. This study was approved by the Human Research Protection Office at Washington University in St Louis, and written informed consent was obtained from all participants.

As part of standard clinical care, preoperative magnetic resonance (MR) images were acquired, with a Siemens Vision 1.5T scanner, and included two T₂-weighted turbo spin-echo sequences: one acquired in transverse planes (TR=8,904 msec; TE=90 msec; flip angle: 180; 53 planes, 1 × 1 × 2-mm voxels), and one acquired in coronal planes (TR: 3,700 msec; TE: 96 msec; flip angle: 180; 19 slices; 1 × 1 × 2-mm voxels). Head movement was prevented during MR imaging by a Leksell stereotactic frame anchored to the skull. Postoperative computed tomography (CT) images were acquired after removal of the stereotactic frame with one of three Siemens Somatom scanners (Definition 64, Sensation 64, or Plus 4), with 120 kV, 206–320 mAs and 0.5 × 0.5 × 1-mm or 0.5 × 0.5 × 2-mm voxels. CT images were examined for movement, recognizable by discontinuities along the skull in coronal and sagittal views. CT scans with significant movement were excluded.

Image-processing and atlas-registration procedures were performed, as described previously.²⁴ Briefly, fiducial structures identified on the MR image and the electrode position from the CT image were used to localize the position of each contact in Mai atlas space.²⁷ Based on the electrode configuration and stimulator settings used in our study, it is likely that current activates neurons and axons within about a 2-mm radius from the center of the contact.²⁸ Thus, a reasonable approximation of the likely suprathreshold effect of monopolar stimulation at a given contact is within a 2-mm radius sphere centered on the contact itself. Bipolar or multipolar settings could alter the pattern and extent of current spread and, thereby, which structures and pathways are affected; therefore, only participants with monopolar settings were included in these analyses.

We used the clinically-determined optimal electrode contact and settings to test the effects of stimulation on mood response (Table 1).

Participants had PD medications withheld overnight and were in the “practical-defined off state”²⁹ at the time of testing. Participants completed motor and mood assessments in four conditions on the same day: 1) bilateral off (DBS OFF); 2) only left unilateral stimulation (left DBS); 3) only right unilateral stimulation (right DBS); and 4) bilateral stimulation (bilateral DBS). Data were collected in a counterbalanced, double-blind manner. Participants were studied at least 42 minutes after changing stimulator conditions to ensure near steady-state motor status during testing.^30,31 Motor signs were measured with the Unified Parkinson Disease Rating Scale (UPDRS) motor subscale by a blinded, validated Movement Disorders specialist who had undergone training and validation requiring intraclass correlations >0.90, as compared with senior investigators trained in UPDRS. Mood response was measured with a computerized version of the Visual Analog Scale (VAS), which is a valid and reliable self-report measure of current mood state³² that has been previously used in studies on PD.^21,33 For each stimulation condition, participants were presented with two opposite adjectives and asked to mark how they felt at that moment on that dimension. The result is quantified as the distance along a horizontal line between the two adjectives (range: 0–100 mm). Adjectives were chosen based on a Circumplex Model of emotion,³⁴ which posits that emotional states can be represented by a two-dimensional circle, in which the proximity to the circumference indicates the intensity of that emotion, and the location around the circle indicates which emotions are related to or opposite from each other. The two primary dimensions in this model are valence (the pleasant or unpleasantness of an emotional state; also referred to as positive and negative affect) and emotional arousal (from low arousal to high arousal). The items included in the VAS were chosen to sample the following operationally-defined mood states: valence (average of responses to sad/happy and grouchy/cheerful items), emotional arousal (average of tranquil/intense, and passive/aroused items), apathy (motivated/apathetic item) and anxiety (average of responses to calm/nervous, relaxed/distressed, and calm/tense items).

To determine whether stimulation condition affected mood response overall, we ran a repeated-measures multivariate analysis of variance (MANOVA), with stimulation condition as the repeated measure and each of the mood-state variables (valence, emotional arousal, anxiety, and apathy) as the multiple measures. Based on the results of this omnibus test, univariate repeated-measures ANOVAs were used to determine which mood states were affected by stimulation. One-sample t-tests were used to determine whether stimulation-induced change in mood response from DBS OFF was significantly different from 0. Chi-square tests were used to assess the proportion of participants reporting DBS-induced improved mood versus those reporting no improvement (no change or a decline in mood). To determine possible laterality effects of stimulation, paired t-tests were used to compare mood ratings across unilateral stimulation conditions. Pearson correlations were used to test the relationship between mood response and contact location (Mai Atlas coordinates, separately along each axis). Because of the rank-order nature of the UPDRS motor ratings, nonparametric analyses (Wilcoxon signed-ranks test, Spearman rank correlations) were used for UPDRS motor ratings. The threshold for statistical significance was set at p <0.05.

Only data from participants with adequate preoperative and postoperative MR and CT scans for atlas registration and bilateral monopolar settings were analyzed. Of the 42 participants who completed mood ratings, we excluded the following: 11 did not have contact location information because of inadequate neuroimaging data or poor atlas-registration because of enlarged ventricles; 5 had bipolar or multipolar contact settings; and 2 did not complete mood ratings for all four stimulation conditions. Analyses are, therefore, based on 24 participants with identified contact locations, monopolar settings, and complete mood data (see Table 1 for clinical and demographic information).

The majority of clinically optimal contacts resided in and near the dorsal portion of the STN and extended into the zona incerta. There was more variability in location along the anterior–posterior and medial–lateral axes for the right DBS than left DBS contacts (Figure 1). Stimulation variables are summarized in Table 1.

One participant reported extreme improvements (>200%) in valence ratings with stimulation and was removed from all analyses of valence ratings as a statistical outlier. Of note, this participant's extreme mood improvement was driven by very low mood ratings in the DBS OFF condition and elevated mood ratings in the stimulation ON conditions; the contact locations for this participant were similar to those for other participants.

Overall, mood ratings improved with stimulation, and the majority of participants reported a benefit of stimulation on mood (Table 2). There was a significant effect of stimulation condition on mood response in general (omnibus test; Wilk's λ=0.68; p<0.02). Univariate repeated-measures ANOVAs revealed a significant effect of stimulation condition on all mood ratings (Figure 2) except emotional arousal. Specifically, stimulation significantly increased valence (F [3, 69]=7.20; p<0.01), decreased anxiety (F [3, 69]=6.81; p<0.01), and decreased apathy (F [3, 69]=3.58; p=0.03). In fact, DBS-induced changes (difference from DBS OFF) were significant for each stimulation condition for all of the mood ratings (one-sample t-tests; all ps <0.05) except emotional arousal and apathy ratings for the right-DBS condition. Furthermore, a significant proportion of participants reported improvements in mood with bilateral STN DBS (Table 2; valence: χ²[1]=7.35; p<0.01; N=23), anxiety: χ²[1]=8.17; p<0.01; N=23), apathy: χ²[1]=6.0; p=0.01; and arousal: NS). Similarly, a majority of participants reported improved mood with unilateral DBS (Table 2), but only valence (left-DBS: χ²[1]=7.35; p<0.01; N=23); and anxiety (left-DBS: χ²[1]=4.17; p=0.04; N=24); right-DBS: χ²[1]=8.17; p<0.01; N=24) reached statistical significance.

To evaluate hemispheric differences in mood response, change in self-reported mood from the DBS OFF condition was computed for both unilateral stimulation conditions. Left-DBS induced significantly greater changes in valence than did right-DBS (t [22]=2.38; p=0.03) and trend-level significance for apathy ratings (t [23]=1.83; p=0.08), although there were no significant differences for emotional arousal or anxiety ratings (paired t-tests).

Right-DBS–induced improvements in valence ratings were related to the “Z” coordinates (r=0.44; p=0.04), such that more dorsal contacts were associated with greater improvements in valence ratings. There was a nonsignificant trend for right-DBS–induced improvement in valence to relate to the “X” coordinate of the contact location (r = –0.39; p=0.07), such that more medial contacts were associated with greater improvement in valence ratings. There were no significant relationships between left active contact location (X, Y, Z coordinates) and the change in mood response to left DBS.

Unilateral and bilateral stimulation improved motor functioning (Table 2). There was a significant improvement in UPDRS motor score with stimulation, as compared with DBS OFF (Wilcoxon signed-ranks test; left-DBS: z = –3.48; p<0.01; right-DBS: z = –4.02; p<0.01; bilateral DBS: z = –4.29; p<0.01). Contralateral motor responses to left-DBS were not different from right-DBS.

There were no significant relationships between contact location (X, Y, and Z coordinates) and the stimulation-induced changes in contralateral UPDRS motor ratings (either left- or right-DBS).

The magnitude of UPDRS motor score improvement (i.e., change from DBS OFF) did not correlate with the magnitude of improvement in mood ratings for either unilateral stimulation condition. For the bilateral DBS condition, the magnitude of UPDRS motor score change correlated with the magnitude of the change in apathy ratings (r_s=0.51; p=0.01) but not with other mood ratings. Furthermore, the only significant relationship between baseline clinical characteristics (age, duration of PD, months since DBS surgery, and OFF/OFF UPDRS motor score) and DBS-induced mood response was between change in arousal ratings and months since DBS surgery (r = –0.44; p=0.04). Thus, there were no consistent relationships between mood response and motor functioning or other clinical characteristics.