Cognitive outcomes following subthalamic nucleus deep brain stimulation (STN DBS) for Parkinson disease (PD) have been mixed; however, numerous studies have observed a moderate decline in verbal fluency (VF).1 The underlying mechanism of VF decline is likely to be multifactorial and is not entirely clear, but anatomical electrode placement has been particularly implicated.2 The impact of stimulation location on VF decline has been evaluated in different STN subregions (active contact localization).3–5 Findings from our previous work indicate that greater anterior displacement of the electrode from the left STN midpoint contributes to greater phonemic VF decline.6 Recently, attention has also focused upon lesioning effects on subcortical and cortical structures following STN DBS, and several studies support an important role for lesioning on VF decline.7–11 However, results are inconsistent and poorly understood due to limited understanding of such structures. In addition, the role of the frontal trajectory and active contact location on VF decline has not been independently examined.8 Furthermore, despite an implicated role for the caudate nucleus on cognitive processes, there are few and inconsistent data on the effects of caudate nucleus penetration on VF decline following STN stimulation in patients with PD.8,12–16 The present study aims to examine the effect of STN DBS electrode trajectory, particularly frontal lobe and caudate lesioning, on VF decline.
Methods
Patient Population and Motor Outcomes
Ethical approval was obtained from the University of Michigan IRB, and informed consent was obtained from all patients. Data were prospectively collected and retrospectively analyzed from adult patients with PD who underwent bilateral STN DBS between 2008 and 2018. Our patient selection criteria have been previously discussed in detail.17
Motor improvement was calculated as the percentage change between scores on the Movement Disorder Society revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS), Section III, in the off-medication/off-stimulation and off-medication/on-stimulation modes at 6 or 12 months following electrode insertion.17
Neuropsychological Evaluation
Patients completed neuropsychological evaluation approximately 3 months preoperatively in the on-medication state and 9 months postoperatively in the on-medication and on-stimulation state. Preoperative and postoperative neuropsychological evaluations included phonemic VF from the Controlled Oral Word Association test18 and semantic VF from the category fluency test.19 In the Controlled Oral Word Association test phonemic VF task, patients were instructed to list as many different words as they could in 1 minute to a letter cue across three trials (“C,” “F,” and “L”). The phonemic VF raw score was computed as the sum of all three trials. In the category fluency test semantic VF task, patients were instructed to list as many different animals as they could in 1 minute. Serving as the primary outcome variables, percent changes in pre- versus postoperative phonemic and semantic VF scores were calculated. Significant VF decline was defined as exceeding 1 SD, which was greater than or equal to approximately a 25% decline, per clinical criteria.20
Surgical Procedure
The MRI protocol, atlas-based targeting, and surgical procedure were as previously described.4,21 Prior to surgery, the planned trajectory was determined by visualization of the STN on 3-T MRI.22 The entry point was chosen to lie in the coronal suture to minimize the risk of epidural hematoma and 2–3 cm off the midline to avoid the sagittal sinus. The final trajectory was selected based on microelectrode recordings. Targeting ensured that the electrode avoided the superior frontal sulcus and lateral ventricle wall.
Image Analysis
Two preoperative MR images were obtained: a 3-T MRI sequence before the day of surgery and a 1.5-T MRI sequence with a Leksell stereotactic frame on the day of surgery. All images were resampled and transformed to Talairach orientation and then coregistered using a mutual-information algorithm with cubic-spline interpolation (0.5 × 0.5 × 0.5–mm voxels; AnalyzeDirect, Inc.). A postoperative CT scan was obtained 2–4 weeks after surgery to visualize DBS electrode contacts after brain shift and intracranial air had resolved, and was coregistered to preoperative MRI sequences. The electrode path was visualized through the subcortical structures, noting whether the path crossed the caudate nucleus, and that was binarily coded as yes or no (Fig. 1).
Coregistered preoperative MR and postoperative CT images showing caudate nucleus penetration in coronal (A) and sagittal (B) projections.
To determine the frontal trajectory of the electrode in 3D space (Fig. 2), the distance of the electrode from midline (ab line) was measured in the coronal projection between the falx cerebri (point a) and the medial border of the electrode at the site of brain entry (point b). The superior frontal gyrus (SFG) width (ac line) was measured between the falx cerebri (point a) and the superior frontal sulcus (point c). These lines were drawn at the uppermost aspect of the brain. The distance of the electrode from midline (ab line) was calculated as a proportion of the full width of the SFG (ac line) by ab/ac. The closer this proportion was to 1, the more lateral the electrode was placed in the SFG. A ratio greater than 1 indicated that the electrode traversed the middle frontal gyrus.
The coronal projection of a coregistered preoperative MR and postoperative CT image showing SFG penetration. Point a is the medial frontal lobe surface close to falx cerebri; point b represents the medial border of the electrode at the site of brain entry; and point c is the lateral border of the SFG.
To localize the active contact, first the STN midpoint was defined as the point halfway between the STN oral and caudal poles, which were identified on coronal MRI sequences. Coordinates of the active contacts were then determined and recorded relative to these MR-visualized STN midpoints. Lateral (X), anterior (Y), and dorsal (Z) directions relative to the STN midpoint were defined as positive.6,23
Statistical Analysis
Independent-sample t-tests were conducted to examine the relationship between electrode penetration of the caudate nucleus and percent change in VF. The t-tests were repeated separately in the right and left hemispheres for both phonemic and semantic VF. Pearson’s correlations were conducted to examine relationships between trajectory (ab/ac) and VF percent change (semantic and phonemic VF) in each hemisphere. Pearson’s r effect sizes were interpreted as follows: 0.10 = small, 0.30 = medium, 0.50 = large.24–26 To limit type I error, only variables with medium or large effect sizes were interpreted as potentially meaningful and included in later regression analysis, as appropriate. To rule out potential confounders in the above analyses, Pearson’s correlations were used to examine relationships between VF change and several demographic, disease, and methodological variables. These included age at baseline, years of education, MDS-UPDRS Section III percent change, disease duration at baseline, and time between neuropsychological evaluations. Finally, to evaluate the unique contribution of surgical variables to VF change, multiple regression was conducted, with the addition of anterior active contact location in the left hemisphere model, given that this was previously identified as a significant predictor of phonemic VF decline.37
Results
Demographics
We examined 59 patients with a mean age of 63 years (range 43–78 years) who underwent bilateral STN DBS for PD. One participant was missing phonemic VF data and was excluded from analyses. DBS electrodes that traversed the middle frontal gyrus in 7 total participants were excluded from analysis (2 participants in both hemispheres, 2 in left only, and 5 in right only). The final sample included 56 participants, with 54 included in left hemisphere analyses and 51 in right hemisphere analyses (see full demographics in Table 1).
Demographic and disease variables in 56 patients with PD
| Variable | Value |
|---|---|
| Sex | |
| Male | 41 (73.2%) |
| Female | 15 (26.8%) |
| Education, yrs | 14.53 ± 2.78 |
| Handedness | |
| Rt | 48 (85.7%) |
| Lt | 8 (14.3%) |
| Age, pre-DBS | 63.17 ± 7.37 |
| Disease duration pre-DBS, yrs since Dx | 10.17 ± 4.70 |
| MDS-UPDRS Section III | |
| Pre-DBS | 41.31 ± 14.44 |
| % change* | 39.32 ± 17.90 |
| LED in mg, pre-DBS | 1469.42 ± 716.24 |
Dx = diagnosis; LED = levodopa-equivalent dosage.
Values are expressed as the number of patients (%) or the mean ± SD.
According to the following formula: % change = ([post-DBS − pre-DBS]/pre-DBS) × 100.
Clinical Outcomes
There were statistically significant motor improvements following STN DBS (Table 1). Percent declines from pre- to post-DBS in phonemic and semantic VF were 17.6% ± 17.2% and 9.25% ± 29.6%, respectively (p < 0.01).
Caudate Nucleus Involvement
Independent-sample t-tests revealed no significant differences in phonemic or semantic VF percent change between groups with caudate nucleus involvement versus no caudate involvement, in both the right (phonemic: t(47) = 0.72, p = 0.47; semantic: t(45) = −1.55, p = 0.13) and left (phonemic: t(49) = −1.62, p = 0.11; semantic: t(50) = 0.60, p = 0.55) hemispheres. Therefore, caudate nucleus involvement was excluded from subsequent regression analyses.
Relationships Between VF Change and Trajectory
In the right hemisphere, Pearson’s correlations revealed that there were no significant correlations between phonemic or semantic VF change and frontal trajectory (r < |0.30|, p > 0.05). In the left hemisphere, there was no significant correlation between semantic VF change and frontal trajectory (r(51) = 0.20, p = 0.14). However, Pearson’s correlation revealed a significant inverse correlation between phonemic VF change and trajectory (r(52) = −0.58, p < 0.001), such that more lateral trajectories were associated with greater phonemic fluency decline. Investigating potential confounding relationships, correlations were small and not statistically significant between phonemic VF change and age at baseline, years of education, MDS-UPDRS Section III percent change, disease duration at baseline, or time between neuropsychological evaluations (r < |0.30|, p > 0.05).27
Relative Contributions of Frontal Trajectory and Active Contact Location
To compare the relative effects of frontal trajectory and active contact location in the left hemisphere, standard multiple regression was conducted with phonemic VF percent change as the dependent variable. The overall model was statistically significant (adjusted R2 = 0.31, F(2, 51) = 13.06, p < 0.001), without multicollinearity (variance inflation factor = 1.08, tolerance = 0.92). Trajectory emerged as a statistically significant predictor, explaining 28% of the variance in phonemic fluency change (unstandardized B = −63.91 ± 13.74, β = −0.55, p < 0.001, sr2 = 0.28), whereas anterior active contact location in the left hemisphere was no longer significant (unstandardized B = −0.71 ± 0.96, β = −0.09, p = 0.46, sr2 = 0.006).
Discussion
There are inconsistent data on the effects of cortical and subcortical electrode penetration on VF decline following STN DBS surgery.8,11,28 We compared the roles of active contact location, electrode frontal trajectory, and caudate nucleus penetration on VF. We found that laterality within the left SFG predicts significantly greater phonemic VF decline, and that lateral SFG penetration was a stronger predictor of phonemic VF decline than active contact location.
Role of the Left Frontal Lobe in VF
Encompassing the Broca area, the importance of the left frontal cortex in VF is well established.29 Meta-analysis of VF in patients with cortical brain lesions has revealed that the largest phonemic VF impairment was associated with unilateral left frontal damage, whereas semantic VF was more dependent on temporal lobe integrity.7 In particular, lesions in the left dorsolateral frontal cortex and striatum have been associated with impairments in phonemic VF.10 Stuss et al.10 assessed VF in 74 patients with focal brain lesions in frontal and nonfrontal regions and observed that damage in the left hemisphere affected both semantic and phonemic VF, whereas lesions in the right hemisphere impacted semantic VF more severely. They found that the left dorsolateral frontal cortex, particularly in the case of superior lesions, was an area associated with the greatest VF decline. Given that DBS electrodes penetrate the frontal cortex, we hypothesized that postoperative impairments in VF might be related to the lesioning in this region. Consistent with this hypothesis, our study found that a lateral SFG trajectory in the left hemisphere was associated with greater phonemic VF decline, whereas a right-sided lateral SFG trajectory did not impact VF.
DBS Lesioning and VF Decline
Unlike large brain lesions, DBS allows for the investigation of smaller (1.2-mm-diameter) microlesions, offering insight into the cortical pathways involved in VF. Previous research has explored the anatomical connection between the lateral SFG and Broca area in the inferior frontal gyrus. Kinoshita et al.30 demonstrated white matter association fibers connecting the Broca area and the lateral SFG by dissection of 8 human cadaveric cerebral hemispheres and diffusion tensor imaging of 53 right-handed patients. In both groups the width of the Broca–lateral SFG tract was greater in the left hemisphere, and these differences were statistically significant. Heim et al.29 used functional MRI to assess activation of Brodmann areas (BAs) 44 and 45, comprising the Broca area, in semantic and phonemic VF tasks. Although both semantic and phonemic VF tasks activated both left BA 44 and BA 45, phonemic VF activated BA 44 more strongly than did semantic VF.22 Our findings supported our hypothesis that penetration of the Broca–lateral SFG tract would be associated with greater phonemic VF decline. That is, greater laterality of DBS lead penetration in the lateral SFG appeared to increase the probability of an intersection with the Broca–lateral SFG pathway, resulting in more significant phonemic VF decline. Our study further suggests that this lesioning effect has greater impact on phonemic VF decline than does active contact location. These findings manifested as clinically significant, with phonemic and semantic VF declines (i.e., exceeding 1 SD, or approximately 25% change, per Heaton norms)20 identified in 37.9% and 36.2% of patients, respectively. Clinically significant decline in phonemic VF equated to an average decline in raw scores from 42.10 ± 9.09 to 27.70 ± 7.28, and clinically significant decline in semantic VF equated to an average decline in raw scores from 19.50 ± 3.78 to 13.15 ± 2.94.
Previous studies have investigated the effects of DBS trajectory on phonemic versus semantic VF. Le Goff and colleagues8 observed that anterior trajectory of left hemispheric electrodes increased the risk of decline in semantic VF, and postulated that anterior left trajectory may disrupt corticocortical pathways linking specific frontal and temporal lobe areas.31 This corroborates evidence of temporal injuries showing larger semantic than phonemic deficits.32 Moreover, they observed that more anterior trajectories penetrated the anterior limb of the internal capsule, which may explain declines in semantic VF.8 York et al.11 found that declines in phonemic VF scores were associated with lateral and superior electrode displacement in the left hemisphere and with posterior and superior displacement in the right hemisphere. By contrast, for semantic VF decline, they found no such associations in either hemisphere. Unlike phonemic VF, our study found no correlation between frontal trajectory and semantic VF decline, consistent with previous findings suggesting a relatively greater contribution of frontal trajectory on phonemic than semantic VF. Finally, Costentin and colleagues33 investigated lesioning of white matter tracts following bilateral STN DBS for PD. Such tracts included the frontal aslant, anterior and long segments of the arcuate, inferior longitudinal anterior thalamic, and frontostriatal tracts, many of which have been implicated in VF. However, no correlation was found between tract disconnection and VF change at 6 months.33 Future studies could investigate the role of the Broca–lateral SFG intersection on VF decline exclusively by applying advanced imaging like diffusion tensor imaging.
Effect of Caudate Nucleus Penetration on VF
Regarding subcortical structures, functional MRI studies have suggested a prominent role for the caudate nucleus and left presupplementary motor area in frontal lobe activation during phonemic more than semantic VF.14 The caudate nucleus has been associated with both working memory and executive function—two important aspects of VF performance.12 DBS electrode penetration has been investigated in a number of studies,12,32,34,35 and Isler and colleagues15 observed a transient interaction of caudate nucleus penetration and VF decline at 3 but not 12 months postoperatively. Similarly, several studies found caudate nucleus penetration effects on VF decline.5,13,36 In our study, 48% and 50% of right- and left-sided electrodes, respectively, penetrated the caudate nucleus. No association was found between caudate nucleus penetration and VF outcome in our study.
Limitations
Our study has several limitations. The data were prospectively obtained but retrospectively analyzed. Hence, our findings uncovered correlative relationships and not causation. All patients underwent operation in a single center using a uniform STN DBS targeting strategy. Hence, effects of more lateral DBS trajectories, including relationships in the middle frontal gyrus, were not studied. Finally, all patients were treated with STN DBS. Hence, effects of similar lateral SFG penetration with alternative DBS targets (e.g., globus pallidus internus) were not examined.
Conclusions
In this study, we demonstrated that phonemic VF decline correlated with more lateral left SFG lead trajectories. One explanation of this finding may be microlesioning of the corticocortical Broca–lateral SFG pathway in the left hemisphere. Knowledge of this finding may improve surgical trajectory planning to reduce the impact of STN DBS on VF.
Disclosures
The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.
Author Contributions
Conception and design: Patil, Askari, Lam, Persad. Acquisition of data: all authors. Analysis and interpretation of data: Patil, Askari, Greif, Lam, Maher. Drafting the article: Askari, Greif, Lam. Critically revising the article: Patil, Greif, Lam, Maher, Persad. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Patil. Statistical analysis: Askari, Greif. Administrative/technical/material support: Patil. Study supervision: Patil.
References
- 1↑
Volkmann J, Daniels C, Witt K. Neuropsychiatric effects of subthalamic neurostimulation in Parkinson disease. Nat Rev Neurol. 2010;6(9):487–498.
- 2↑
Højlund A, Petersen MV, Sridharan KS, Østergaard K. Worsening of verbal fluency after deep brain stimulation in Parkinson’s disease: a focused review. Comput Struct Biotechnol J. 2016;15:68–74.
- 3
Floden DP, Matias CM, Wathen CA, Ozinga GE, Hogue O, Machado AG. Contact location and neuropsychological outcomes in subthalamic deep brain stimulation. Neurosurgery. 2018;83(4):666–674.
- 4↑
Mossner JM, Chou KL, Maher AH, Persad CC, Patil PG. Localization of motor and verbal fluency effects in subthalamic DBS for Parkinson’s disease. Parkinsonism Relat Disord. 2020;79:55–59.
- 5↑
Witt K, Granert O, Daniels C, et al. Relation of lead trajectory and electrode position to neuropsychological outcomes of subthalamic neurostimulation in Parkinson’s disease: results from a randomized trial. Brain. 2013;136(Pt 7):2109–2119.
- 6↑
Greif TR, Askari A, Cook Maher A, Patil PG, Persad C. Anterior lead location predicts verbal fluency decline following STN-DBS in Parkinson’s disease. Parkinsonism Relat Disord. 2021;92:36–40.
- 7↑
Henry JD, Crawford JR. A meta-analytic review of verbal fluency performance following focal cortical lesions. Neuropsychology. 2004;18(2):284–295.
- 8↑
Le Goff F, Derrey S, Lefaucheur R, et al. Decline in verbal fluency after subthalamic nucleus deep brain stimulation in Parkinson’s disease: a microlesion effect of the electrode trajectory? J Parkinsons Dis. 2015;5(1):95–104.
- 9
Okun MS, Gallo BV, Mandybur G, et al. Subthalamic deep brain stimulation with a constant-current device in Parkinson’s disease: an open-label randomised controlled trial. Lancet Neurol. 2012;11(2):140–149.
- 10↑
Stuss DT, Alexander MP, Hamer L, et al. The effects of focal anterior and posterior brain lesions on verbal fluency. J Int Neuropsychol Soc. 1998;4(3):265–278.
- 11↑
York MK, Wilde EA, Simpson R, Jankovic J. Relationship between neuropsychological outcome and DBS surgical trajectory and electrode location. J Neurol Sci. 2009;287(1-2):159–171.
- 12↑
Arsalidou M, Duerden EG, Taylor MJ. The centre of the brain: topographical model of motor, cognitive, affective, and somatosensory functions of the basal ganglia. Hum Brain Mapp. 2013;34(11):3031–3054.
- 13↑
Bot M, van den Munckhof P, Schmand BA, de Bie RMA, Schuurman PR. Electrode penetration of the caudate nucleus in deep brain stimulation surgery for Parkinson’s disease. Stereotact Funct Neurosurg. 2018;96(4):223–230.
- 14↑
Grogan A, Green DW, Ali N, Crinion JT, Price CJ. Structural correlates of semantic and phonemic fluency ability in first and second languages. Cereb Cortex. 2009;19(11):2690–2698.
- 15↑
Isler C, Albi A, Schaper FL, Temel Y, Duits A. Neuropsychological outcome in subthalamic nucleus stimulation surgeries with electrodes passing through the caudate nucleus. Stereotact Funct Neurosurg. 2016;94(6):413–420.
- 16
Parsons TD, Rogers SA, Braaten AJ, Woods SP, Tröster AI. Cognitive sequelae of subthalamic nucleus deep brain stimulation in Parkinson’s disease: a meta-analysis. Lancet Neurol. 2006;5(7):578–588.
- 17↑
Chou KL, Taylor JL, Patil PG. The MDS-UPDRS tracks motor and non-motor improvement due to subthalamic nucleus deep brain stimulation in Parkinson disease. Parkinsonism Relat Disord. 2013;19(11):966–969.
- 18↑
Ruff RM, Light RH, Parker SB, Levin HS. Benton Controlled Oral Word Association Test: reliability and updated norms. Arch Clin Neuropsychol. 1996;11(4):329–338.
- 19↑
Morris JC, Heyman A, Mohs RC, et al. The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD). Part I. Clinical and neuropsychological assessment of Alzheimer’s disease. Neurology. 1989;39(9):1159–1165.
- 20↑
Heaton RK, Miller SW, Taylor MJ, Grant I. Revised Comprehensive Norms for an Expanded Halstead-Reitan Battery: Demographically Adjusted Neuropsychological Norms for African American and Caucasian Adults, Professional Manual. Psychological Assessment Resources; 2004.
- 21↑
Conrad EC, Mossner JM, Chou KL, Patil PG. Atlas-independent, electrophysiological mapping of the optimal locus of subthalamic deep brain stimulation for the motor symptoms of Parkinson disease. Stereotact Funct Neurosurg. 2018;96(2):91–99.
- 22↑
Patil PG, Conrad EC, Aldridge JW, Chenevert TL, Chou KL. The anatomical and electrophysiological subthalamic nucleus visualized by 3-T magnetic resonance imaging. Neurosurgery. 2012;71(6):1089–1095.
- 23↑
Houshmand L, Cummings KS, Chou KL, Patil PG. Evaluating indirect subthalamic nucleus targeting with validated 3-tesla magnetic resonance imaging. Stereotact Funct Neurosurg. 2014;92(6):337–345.
- 24
Bakeman R. Recommended effect size statistics for repeated measures designs. Behav Res Methods. 2005;37(3):379–384.
- 25
Cohen J. A power primer. Psychol Bull. 1992;112(1):155–159.
- 26
Field A. Discovering Statistics Using IBM SPSS Statistics. 4th ed. SAGE;2013.
- 27↑
Schneider JS, Sendek S, Yang C. Relationship between motor symptoms, cognition, and demographic characteristics in treated mild/moderate Parkinson’s disease. PLoS One. 2015;10(4):e0123231.
- 28↑
Lefaucheur R, Derrey S, Martinaud O, et al. Early verbal fluency decline after STN implantation: is it a cognitive microlesion effect? J Neurol Sci. 2012;321(1-2):96–99.
- 29↑
Heim S, Eickhoff SB, Amunts K. Specialisation in Broca’s region for semantic, phonological, and syntactic fluency? Neuroimage. 2008;40(3):1362–1368.
- 30↑
Kinoshita M, Shinohara H, Hori O, et al. Association fibers connecting the Broca center and the lateral superior frontal gyrus: a microsurgical and tractographic anatomy. J Neurosurg. 2012;116(2):323–330.
- 31↑
Daniels C, Krack P, Volkmann J, et al. Risk factors for executive dysfunction after subthalamic nucleus stimulation in Parkinson’s disease. Mov Disord. 2010;25(11):1583–1589.
- 32↑
Ellfolk U, Joutsa J, Rinne JO, Parkkola R, Jokinen P, Karrasch M. Striatal volume is related to phonemic verbal fluency but not to semantic or alternating verbal fluency in early Parkinson’s disease. J Neural Transm (Vienna). 2014;121(1):33–40.
- 33↑
Costentin G, Derrey S, Gérardin E, et al. White matter tracts lesions and decline of verbal fluency after deep brain stimulation in Parkinson’s disease. Hum Brain Mapp. 2019;40(9):2561–2570.
- 34↑
Bauer E, Toepper M, Gebhardt H, Gallhofer B, Sammer G. The significance of caudate volume for age-related associative memory decline. Brain Res. 2015;1622:137–148.
- 35↑
Grahn JA, Parkinson JA, Owen AM. The cognitive functions of the caudate nucleus. Prog Neurobiol. 2008;86(3):141–155.
- 36↑
Tesio V, Rizzi L, Jiang T, Fronda C, Lanotte M, Castelli L. Deep brain stimulation of the subthalamic nucleus in Parkinson’s disease: relationship between the electrode trajectory and cognitive decline. Parkinsonism Relat Disord. 2019;61:45–49.
- 37↑
Greif TR, Askari A, Cook Maher A, Patil PG, Persad C. Anterior lead location predicts verbal fluency decline following STN-DBS in Parkinson’s disease. Parkinsonism Relat Disord. 2021;92(36):40.
