Learning from negative consequences is impaired by STN-DBS and levodopa in Parkinson’s disease

Abhinav Raina; Roopa Rajan; Gangadhara Sarma; Syam Krishnan; Krishnakumar Kesavapisharady; Asha Kishore

doi:10.4103/AOMD.AOMD_54_20

ORIGINAL ARTICLES

Year : 2021 | Volume : 4 | Issue : 2 | Page : 66-72

Learning from negative consequences is impaired by STN-DBS and levodopa in Parkinson’s disease

Abhinav Raina¹, Roopa Rajan², Gangadhara Sarma³, Syam Krishnan³, Krishnakumar Kesavapisharady³, Asha Kishore⁴
¹ Department of Neurology, Birla Institute of Medical Research, Gwalior, Madhya Pradesh, India
² Department of Neurology, All India Institute of Medical Sciences, New Delhi, India
³ Comprehensive Care Centre for Movement Disorders, Sree Chitra Tirunal Institute for Medical Sciences, Trivandrum, Kerala, India
⁴ Centre of Excellence in Neurosciences, Aster Medicity, Kochi, Kerala, India

Date of Submission	11-Oct-2020
Date of Decision	27-Mar-2021
Date of Acceptance	19-Apr-2021
Date of Web Publication	11-Jun-2021

Correspondence Address:
Asha Kishore
Centre of Excellence in Neurosciences, Aster Medicity, Kochi, Kerala.
India

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/AOMD.AOMD_54_20

Abstract

CONTEXT: Subthalamic nucleus deep brain stimulation (STN-DBS) and levodopa therapy are reported to produce impulsivity in PD. We tested the hypothesis that STN-DBS enhances impulsive decision-making and that this effect is masked by the beneficial influence of the concomitant reduction in levodopa therapy. AIMS: To compare learning from negative consequences in patients with PD receiving STN-DBS to those without surgery. SETTINGS AND DESIGN: We conducted a prospective study in the Movement Disorder Clinic of a tertiary care university hospital in India, recruiting 26 PD patients treated with STN-DBS (PD-DBS) and 25 matched control patients on medical therapy (PD-MED) alone. METHODS: Impulsivity and decision-making were assessed using the Iowa Gambling Task (IGT) at baseline and 3 months. Dopamine agonist dose was unchanged during the study period. IGT total and block scores were used to assess impulsive decision-making and task learning. STATISTICAL ANALYSIS: General linear mixed model involving repeated measures ANOVA was used to compare IGT total and block scores. RESULTS: There was no independent effect of STN-DBS on the IGT total score (P = 0.198). In the PD-DBS group, block scores worsened after surgery with a negative slope across blocks, suggesting a lack of task learning [TIME*BLOCK*INTERVENTION [F (4, 46) = 4.810, P = 0.003, partial Eta-squared = 0.095]. In the PD-MED group, block scores were seen to improve from baseline and across the blocks. These contrasting effects remained robust when tested in the Drug ON or OFF states. In the PD-DBS group, better task learning and higher block 5 scores were associated with greater post-operative reductions in total LEDD [TIME*BLOCK*LEDD [F (4, 46) = 3.818, P = 0.012, partial Eta-squared = 0.077]. CONCLUSIONS: STN-DBS did not independently affect a global measure of decision making. However, STN-DBS and dopaminergic medications were shown to exert subtle yet significant opposing effects on the ability to learn from negative consequences.

Keywords: Decision-making, impulsivity, Iowa Gambling Task, learning, Parkinson’s disease, STN-DBS

How to cite this article:
Raina A, Rajan R, Sarma G, Krishnan S, Kesavapisharady K, Kishore A. Learning from negative consequences is impaired by STN-DBS and levodopa in Parkinson’s disease. Ann Mov Disord 2021;4:66-72

How to cite this URL:
Raina A, Rajan R, Sarma G, Krishnan S, Kesavapisharady K, Kishore A. Learning from negative consequences is impaired by STN-DBS and levodopa in Parkinson’s disease. Ann Mov Disord [serial online] 2021 [cited 2023 Mar 23];4:66-72. Available from: https://www.aomd.in/text.asp?2021/4/2/66/318093

FNx01Abhinav Raina and Roopa Rajan contributed equally to this manuscript.

Key Messages:

Levodopa and STN-DBS exert subtle and opposing effects on decision making.

Introduction

Impulsivity is a multifactorial construct that leads to inaptly deliberated actions or decisions, without having adequate forethought or consideration of the consequences.^[1] In Parkinson’s Disease (PD), pathologically increased impulsivity manifests as impulse control disorders (ICD) in 2–31% of patients.^[2],[3],[4] PD patients without ICD also demonstrate impulsivity in decision making and response inhibition tasks such as the Iowa Gambling Task (IGT) and the GO NO GO task.^[5],[6],[7]

Subthalamic Nucleus Deep Brain Stimulation (STN-DBS) is the standard of care for patients with PD who experience incomplete relief of motor symptoms with dopaminergic drugs or disabling side effects of these drugs.^[8] The STN is also implicated in modulation of response inhibition and impulsivity.^[9] It is thought that the STN sends a “hold your horses” or “no go” signal to temporarily raise the response threshold and allow time for information accumulation before a decision is made and a response is produced.^[10] STN-DBS is proposed to interfere with this normal function of the STN in situations of conflict, resulting in fast impulsive responses.^[10] This view is supported by the de-novo occurrence of ICD in a subset of patients in the initial years after STN-DBS.^[11] In addition, STN-DBS has been shown to have heterogenous effects on different types of learning—STN-DBS facilitated active feedback learning while impairing working memory and enhancing proactive interference during new learning.^{[12],[13],[14]} On the contrary, pre-existing clinical ICD and impulsivity may improve, concomitant with the reduction in dopaminergic medications following STN-DBS.^[15] Acute lesioning effects, mood and motivation which can potentially affect task performance may also underlie these contrasting effects.^[16]

We sought to test the hypothesis that STN-DBS worsens impulsive decision making and this side effect is masked by the beneficial influence of the concomitant reduction in levodopa and dopamine agonist doses on impulsivity. Multiple lines of evidence implicate dopamine agonists to be among the strongest risk factors for development of ICD and withdrawal of these drugs often results in resolution of clinical impulsivity.^[3],[4] Levodopa reduction being unavoidable after STN-DBS, we eliminated the potential confounding effect of dopamine agonist dose reduction by keeping its dose stable during the study period. To minimize the acute effects of microlesioning, we assessed the patients 3 months post-surgery. We used the IGT to compare the performance of DBS patients with a matched group of medically managed PD during the study period.

Subjects and Methods

This prospective study recruited PD patients from the Movement Disorders clinic of a tertiary care university hospital that has two decades of experience in performing STN-DBS. All subjects gave written informed consent and the Ethics Committee of the hospital approved the study.

Subjects

We recruited PD patients aged >30 years, with at least 4 years duration of symptoms who experienced troublesome motor fluctuations and dyskinesias from June 2015 to December 2017. PD was diagnosed by experienced movement disorder specialists using the UKPDS brain bank criteria.^[17] Consecutive patients selected for STN-DBS and subsequently underwent surgery formed the PD-DBS group. Age, gender, and disease duration-matched patients awaiting STN-DBS were chosen from the surgical waiting list as the medically treated group (PD-MED). Patients with pre-existing ICD (modified Minnesota Impulsive Disorders Interview), Mini Mental State Examination (MMSE) score <24, and Becks depression inventory (BDI) score of >25 were excluded.^[18],[19]

Data collection

MMSE (for cognitive ability), Beck’s Depression Inventory (for depression), modified Minnesota Impulsive Disorders Interview (mMIDI, for ICDs), Barratt Impulsivity Scale, and Apathy Scale (for apathy) were administered during the screening visit of the study, 1h after the patient took the routine morning medications (Drug ON).^{[18],[19],[20],[21],[22]} Clinical data pertaining to age of onset, duration of illness, details of drug therapy, clinical examination, and UPDRS scores were collected. Levodopa equivalent daily dosages of drugs (LEDD) were calculated according to published formula.^[23] A computerized version of the IGT (Millisecond Software, LLC, Washington, USA) was randomly administered after 12-h (overnight) withdrawal of all dopaminergic drugs, (Drug OFF) or 1h after the usual dose of dopaminergic drugs (Drug ON), with a 1 week gap between the two assessments. The PD-DBS group underwent MRI-guided, bilateral STN-DBS with 5-channel microelectrode recording and macrostimulation for target localization as previously reported.^[24] The PD-MED group was continued on stable doses of medical therapy. Three months after surgery, the PD-DBS group performed IGT again in two conditions: (1) Stimulator ON-Drug ON; (2) Stimulator ON-Drug OFF. In the PD-MED group, IGT was repeated after 3 months both in the Drug ON and Drug OFF states. Dose of dopamine agonist was held constant during the study period. BDI, Apathy scale, and mMIDI were also administered at the 3-month follow up, in the Drug ON state.

Iowa Gambling Task

The IGT is a validated measure of decision making in which subjects have to maximize profits and avoid losses while drawing cards one by one from four identical decks.^[25] Subjects were allowed 100 draws. Decks “A” and “B” were net disadvantageous, whereas decks “C” and “D” were net advantageous. Decks “A” and “C” had a low frequency of gain (frequent punishments) while decks “B” and “D” had a high frequency of gain (infrequent punishments). Subjects were instructed to earn as much money as possible. Total score was calculated as number of draws from advantageous decks minus number of draws from disadvantageous decks. Block scores were calculated for each block of 20 draws to analyze task learning. For the total IGT score and block scores, values could range from negative to positive, with lower values and negative values indicating worse performance (more impulsive decision making). The construct validity of the original IGT metric (total score) is considered reasonable as a measure of the emotional decision-making process, with the caveat that personality traits and state mood may affect the performance.^[25]

Statistical analysis

The primary outcome measure was the total IGT score at 3 months. We assessed the IGT block scores, a measure of task learning, as a secondary outcome. A general linear mixed model involving repeated measures ANOVA was used to compare the total IGT score with TIME (Baseline, 3 months) and DRUG (Drug ON, Drug OFF) as within-subject factors and INTERVENTION (PD-DBS, PD-MED) as the between-subjects factor. A second model incorporating BLOCK as an additional within-subjects factor was used to compare the IGT block scores. Change in LEDD at 3 months was incorporated as a co-variate in both the models. Additionally, post-hoc sensitivity analysis was performed incorporating baseline LEDD as a co-variate in both the models, as statistically significant changes were noted in this variable at baseline. All statistical analyses were performed using the SPSS statistical software package (release 16.0, SPSS Inc.; Chicago, IL).

Results

A total of 51 subjects were enrolled with 26 in the PD-DBS group and 25 in the PD-MED group. Their baseline characteristics are described in [Table 1]. Both groups were matched for all baseline parameters except the total LEDD, which was higher in the PD-DBS group (P = 0.002). There was a reduction in total LEDD of 39.6 ± 0.2% at 3 months in the PD-DBS group, while the total LEDD remained the same in the PD-MED group. There was no change in the dose of dopamine agonist in either group. Addenbrooke’s cognitive examination (P = 0.09) and Trail Making Test (P = 0.46) scores were similar before and after STN-DBS. Minimal deterioration in verbal fluency (pre-DBS: 12.0 ± 1.9, post-DBS: 11.0 ± 1.9, P = 0.001) was seen post-DBS, though this was not clinically significant. Depression (P = 0.374) and apathy (P = 0.255) scores did not change significantly at 3 months. None of the patients developed new onset ICD at 3 months.

Table 1: Baseline characteristics of the study population

Click here to view

Iowa Gambling Task

Mean IGT total scores are shown in [Table 2] and block scores in [Table 3]. There was no significant main effect of INTERVENTION (STN-DBS) on the IGT total score [F (1, 43) = 1.703, P = 0.198] [Figure 1]. Within the PD-DBS and PD-MED groups, TIME and DRUG (Drug ON/OFF status) did not have a significant effect on the IGT total score [TIME F (1, 47) = 3.410, P = 0.071; DRUG F (1, 47) = 0.012, P = 0.914]. Adjusting for baseline LEDD did not alter the results [INTERVENTION F (1,47) = 3.072, P = 0.086; TIME F (1, 47) = 0.549, P = 0.462; DRUG F (1, 47) = 0.574, P = 0.453] significantly. No statistically significant interactions were noted for INTERVENTION, TIME, DRUG, or LEDD on the total IGT score.

Table 2: IGT total scores in PD-DBS and PD-MED groups

Click here to view

Table 3: Unadjusted absolute IGT block scores in PD-DBS and PD-MED groups, at baseline and at 3 months, in drug ON and OFF states

Click here to view

Figure 1: IGT Total scores in PD-DBS (1) and PD-MED (2) groups at baseline and follow up. Scores in the PD-DBS group showed a declining trend (worsening), though it was not statistically significant. Covariate in the model, LEDD change = 0.202

Click here to view

In the model assessing IGT block scores, there was no significant independent overall effect of INTERVENTION [F (4, 46) = 1.750, P = 0.192]. For the overall model, there were no significant within-subject effects for BLOCK [F (4, 46) = 1.097, P = 0.348] and TIME [F (4, 46) = 3.780, P = 0.058] suggesting poor task learning both across the blocks during the task and while re-testing at 3 months. Results remained similar after adjusting for baseline LEDD [INTERVENTION F (4, 46) = 3.187, P = 0.81; TIME F (4, 46) = 0.236, P = 0.630; BLOCK F (4, 46) = 0.947, P = 0.438]. However, significant within-subject interactions were noted for the interactions among TIME*BLOCK [F (4, 46) = 3.951, P = 0.010, partial Eta-squared = 0.079], TIME*BLOCK*LEDD [F (4, 46) = 3.818, P = 0.012, partial Eta-squared = 0.077] and TIME*BLOCK*INTERVENTION [F (4, 46) = 4.810, P = 0.003, partial Eta-squared = 0.095]. The within-subject interactions were insignificant when DRUG (Drug ON or Drug OFF) was additionally incorporated into the model; TIME*DRUG*BLOCK*INTERVENTION [F (4, 46) = 0.365, P = 0.805]. The significant interactions are further explored in [Figure 2] and [Figure 3]. In the PD-DBS group, the block scores worsened at 3 months compared with baseline and the trend of scores across blocks showed a negative slope, suggesting a lack of learning and worsening performance. This effect was consistent in both Drug ON and Drug OFF testing [Figure 2]A]. In contrast, in the PD-MED group, block scores were better at 3 months and showed a positive slope across the blocks, suggesting task learning. These results were also consistent while patient were tested in Drug ON and Drug OFF states [[Figure 2]B].

Figure 2: Estimated marginal means of the IGT block score change from baseline at DRUG OFF and DRUG ON in the PD-DBS group (A) and PD-MED group (B). In the PD-DBS group, note the decreasing values across blocks and actual values in the negative, suggesting a lack of learning across blocks and worsening performance. In contrast, in the PD-MED group values increase across the blocks with positive individual values, suggesting learning of the task across blocks and improving performance. Covariate LEDD change was evaluated at LEDD change = 0.196

Click here to view

Figure 3: Estimated marginal means of IGT Block scores across the spectrum of dopaminergic medication reduction (LEDD) post-surgery, while tested in stimulation ON, DRUG OFF status (A) and stimulation ON, DRUG ON status (B). The legend shows LEDD change post-surgery. LEDD: Levodopa Equivalent Daily Dosage

Click here to view

To explore the effect of LEDD change on IGT block scores, we classified the LEDD reduction post STN-DBS into five categories, 0% (n = 2), 1–30% (n = 4), 31–50% (n = 11), 51–70% (n = 8), and ≥71% (n = 0). Three months after STN-DBS, effects of LEDD reduction were most pronounced when patients were tested in Drug OFF [Figure 3A]. In DRUG OFF, a clear trend towards better learning of the task and higher block 5 scores were seen with greater reductions in LEDD. Subjects who had no LEDD reduction performed the worst and those with LEDD reduction 51–70% performed the best. When the test was repeated in DRUG ON, overall performance worsened and the gradient effects of LEDD reduction were not prominent [Figure 3B].

Discussion

The present study assessed the impact of bilateral STN-DBS on impulsive decision making using IGT at 3 months post-surgery in advanced PD patients. Following STN DBS, the IGT total scores did not worsen from baseline either in the Drug ON or Drug OFF states. This could mean that either STN-DBS has no independent negative effect on decision making in PD patients or that, IGT total score alone is an inadequate measure of impulsivity. Previous studies have proposed that more sensitive markers such as block score changes and deck choice analysis may be more appropriate.^[5],[26] Moreover, IGT total scores exhibit wide variability and outlier values, which affect the results. This is reflected in our results as well; a large range of IGT total score values ranging from negative to positive were encountered which complicated statistical analysis and assumptions regarding reliability of results. Therefore, in addition to total IGT scores, we pre-specified the analysis of block scores to analyze the task learning ability (choosing favorable decks by learning the negative consequences of picking the wrong decks in early blocks) besides the overall assessment of impulsive decision making from the total scores.

In the PD-DBS group, the block score analysis revealed that task learning was absent after surgery, as shown by decreasing score changes across the blocks both in drug ON and OFF states. This effect of STN-DBS remained significant even after adjusting for the reduction in LEDD. STN-DBS thus appears to interfere with the basal ganglia circuit function of providing the no-go signal during decision making and hampers the initial step of learning from the negative consequences. This is consistent with the evidence from computational models that the STN plays a critical role in the cortico-ventral striatal circuitry responsible for decision making by providing breaking threshold.^[10]

Analysis of the interaction between LEDD reduction and block scores revealed a consistent trend towards better learning in patients who had a greater post-operative reduction in LEDD. These results are consistent with previous studies which reported an improvement in impulsivity with profound medication reduction after STN-DBS.^[15] The effect of LEDD reduction was most pronounced when the subjects were tested in the Drug OFF indicating a long and short duration benefit of levodopa withdrawal despite the overall net negative effect of stimulation. In the drug ON state, the gradient effect was no longer apparent and the overall performance worsened, suggesting that acute dopaminergic boost may also contribute to the defective learning post-DBS. This is corroborated by previous studies that reported poor performance on IGT in PD patients in the Drug ON state.^[27] Our results suggest that a complex interplay of factors may determine net impulsivity after STN-DBS. In this group of subjects, the significant interaction of STN-DBS was directed towards worsened learning, whereas post-operative LEDD reduction tended to improve learning. In addition, acute dopaminergic drug dosing tended to mask the directional effects of both.

In contrast to previous reports that tested the effect of STN-DBS soon after DBS, in this study the IGT was performed at 3 months post-surgery, when the acute micro-lesioning effect of electrode passages is expected to have improved.^[28] In addition those with pre-existing hyperdopaminergic behaviors were excluded and dopamine agonist treatment was kept unaltered in both groups during the study period, both of which could also hinder assessing the independent assessment of the effects of STB-DBS on impulsive decision making. Variations in the tasks used to assess impulsivity, such as the game of dice task that primarily assesses decision making under risk (and is understood to more closely parallel executive functions than the IGT) or delay discounting tasks instead of the IGT that assesses decision making under uncertainty may explain the difference in our results from previous reports of improvement in impulsivity after STN-DBS.^{[28],[29],[30]} Furthermore, variations in study design across cross-sectional and prospective studies maybe contributing to the observed heterogeneity in results.^[27],[30] These methodological differences, while adding to our understanding of a variety of decision making processes in heterogenous settings after STN-DBS, make a direct comparison of results across studies difficult. There was no clinically relevant deterioration in general measures of cognition and executive function after STN-DBS in our study, suggesting that deficits in IGT did not occur in the context of global cognitive dysfunction.

The strength of our study is the study design to control for potential confounders that affect impulsivity and performance in neuropsychological tasks such as exclusion of patients with pre-existing diagnosis of ICD, depression, and apathy and maintenance of a stable dose of dopamine agonist drugs in both medically and surgically treated patients during the study period.^[31] Apathy, reduced motivation, and depression, which are known to follow STN-DBS, were measured via self-assessment questionnaire methods at baseline and at the end of the study.^[32] In the current study, no significant differences were noted in the apathy or BDI scores thereby excluding that such effects of STN DBS could have affected the performance at 3 months. We also demonstrated that there was no learning effect of repeated administration of the IGT on the total IGT score in the medically treated group (twice at visit 1 and twice after 3 months in medication ON and OFF state). This is consistent with previous reports.^[33] We suspect that edema and micro-lesioning effect may not be responsible for the impairment seen in learning the task in the surgical group, as most of such an effect is expected to subside in 3 months. However, we did not test these subjects after surgery in the stimulation OFF-Drug OFF condition to exclude the impact of any persistent micro-lesioning effect.

A limitation of our study is that post-operative imaging was not available to identify the exact location of active contacts within the STN, which could account for some variability in performance. Specifically, stimulation via ventrally located electrodes may be expected to have a greater effect on cognitive and limbic functions compared with dorsally located electrodes.^[34] However, the substantial improvement in UPDRS III motor scores in the same patients is indirectly indicative of good placement of the electrode. Although this study includes a larger number of patients compared with previous reports, given the variability in the total IGT scores, future studies including larger number of patients may be required to confirm these findings.

Bilateral STN-DBS and dopaminergic medications in advanced PD exert subtle yet significant effects on impulsive decision making as assessed by the block score analysis of performance in IGT. STN-DBS impaired task learning and affected the ability to change strategies based on the negative consequences of wrong choices made which resulted in continued unfavorable choices and deteriorating performance. The beneficial effect on decision making of chronic reduction in non-physiological doses of dopaminergic treatment following surgery may mask the adverse effect of STN-DBS on impulsivity and escape detection when tested with tasks such as IGT, especially in the drug ON state. Future studies including larger number of patients and more sensitive tests may be required to confirm the independent effect of STN-DBS on impulsive decision making.

Acknowledgement

The authors express their gratitude to all the patients who took part in the study.

Author Contributions

Raina A: Organization, Execution, Writing of the first draft

Rajan R: Conception, Organization, Design, Execution, Writing of the first draft, Review and Critique

Sarma G: Execution, Review and Critique, Review and Critique

Krishnan S: Organization, Execution, Writing of the first draft, Review and Critique

Kesavapisahrady K: Organization, Execution, Review and Critique

Kishore A: Conception, Organization, Design, Review and Critique

Ethical Compliance Statement

This study was approved by the Institutional Ethics Committee and written informed consent was obtained from all participants.

Conflicts of interest

There are no conflicts of interest.

Financial support and sponsorship

Nil.

References

1.	Madden GJ, Johnson PS. A delay-discounting primer. In Madden GJ, Bickel WK, editors. Impulsivity: The Behavioral and Neurological Science of Discounting. Washington, DC: American Psychological Association; 2010. pp. 11-37.
2.	Voon V, Hassan K, Zurowski M, de Souza M, Thomsen T, Fox S, et al. Prevalence of repetitive and reward seeking behaviors in Parkinson disease. Neurology 2006;67:1254-7.
3.	Weintraub D, Koester J, Potenza MN, Siderowf AD, Stacy M, Voon V, et al. Impulse control disorders in Parkinson disease: A cross-sectional study of 3090 patients. Arch Neurol 2010;67:589-95.
4.	Sarathchandran P, Soman S, Sarma G, Krishnan S, Kishore A. Impulse control disorders and related behaviors in Indian patients with Parkinson’s disease. Mov Disord 2013;28: 1901-2.
5.	Kobayakawa M, Koyama S, Mimura M, Kawamura M. Decision making in Parkinson’s disease: Analysis of behavioral and physiological patterns in the Iowa gambling task. Mov Disord 2008;23:547-52.
6.	Poletti M, Frosini D, Lucetti C, Del Dotto P, Ceravolo R, Bonuccelli U. Decision making in de novo Parkinson’s disease. Mov Disord 2010;25:1432-6.
7.	Manza P, Amandola M, Tatineni V, Li CR, Leung HC. Response inhibition in Parkinson’s disease: A meta-analysis of dopaminergic medication and disease duration effects. NPJ Parkinsons Dis 2017;3:23.
8.	Deuschl G, Schade-Brittinger C, Krack P, Volkmann J, Schäfer H, Bötzel K, et al. A randomized trial of deep-brain stimulation for Parkinson’s disease. N Engl J Med 2006;355: 896-908.
9.	Hershey T, Revilla FJ, Wernle A, Gibson PS, Dowling JL, Perlmutter JS. Stimulation of STN impairs aspects of cognitive control in PD. Neurology 2004;62:1110-4.
10.	Frank MJ, Samanta J, Moustafa AA, Sherman SJ. Hold your horses: Impulsivity, deep brain stimulation, and medication in Parkinsonism. Science 2007;318:1309-12.
11.	Kim A, Kim YE, Kim HJ, Yun JY, Yang HJ, Lee WW, et al. A 7-year observation of the effect of subthalamic deep brain stimulation on impulse control disorder in patients with Parkinson’s disease. Parkinsonism Relat Disord 2018;56:3-8. doi: 10.1016/j.parkreldis.2018.07.010.
12.	Meissner SN, Sudmeyer M, Keitel A, Pollok B, Bellebaum C. Facilitating effects of deep brain stimulation on feedback learning in Parkinson’s disease. Behav Brain Res 2016;313:88-96. doi: 10.1016/j.bbr.2016.06.062
13.	Hershey T, Wu J, Weaver PM, Perantie DC, Karimi M, Tabbal SD, et al. Unilateral vs. Bilateral STN DBS effects on working memory and motor function in Parkinson disease. Exp Neurol 2008;210:402-8. doi: 10.1061/j.expneurol.2007.11.011
14.	Georgiev D, Roskar S, Cus A, Wilkinson L, Jahanshahi M. STN-DBS increases proactive but not retroactive interference during verbal learning in PD. Mov Disord 2020. doi: 10.1002/mds.28423.
15.	Castrioto A, Funkiewiez A, Debû B, Cools R, Lhommée E, Ardouin , et al. Iowa gambling task impairment in Parkinson’s disease can be normalized by reduction of dopaminergic medication after subthalamic stimulation. J Neurol Neurosurg Psychiatry 2015;86:186-90.
16.	Drapier D, Drapier S, Sauleau P, Haegelen C, Raoul S, Biseul I, et al. Does subthalamic nucleus stimulation induce apathy in Parkinson’s disease? J Neurol 2006;253:1083-91. https://doi.org/10.1007/s00415-006-0177-0.
17.	Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of clinical diagnosis of idiopathic Parkinson’s disease: A clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatry 1992;55:181.
18.	Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189-98.
19.	Beck AT, Steer RA, Ball R, Ranieri W. Comparison of beck depression inventories -IA and -II in psychiatric outpatients. J Person Assess 1996;67:588-97.
20.	Christenson GA, Faber RJ, deZwaan M. Compulsive buying: Descriptive characteristics and psychiatric co morbidity. J Clin Psychiatry 1994;55:5-11.
21.	Patton JH, Stanford MS, Barratt ES. Factor structure of the Barratt impulsiveness scale. J Clin Psychol 1995;51:768-77.
22.	Starkstein SE, Mayberg HS, Preziosi TJ, Andrezejewski P, Leiguarda R, Robinson RG. Reliability, validity and clinical correlates of apathy in Parkinson’s disease. J Neuropsychiatry Clin Neurosci1992;4:134-9.
23.	Tomlinson CL, Stowe R, Patel S. Systematic review of levodopa dose equivalency reporting in Parkinson’s disease. Mov Disord 2010;25:2649-85.
24.	Krishnan S, Pisharady KK, Divya KP, Shetty K, Kishore A. Deep brain stimulation for movement disorders. Neurol India 2018;66(Supplement):S90-101.
25.	Bechara A, Damásio AR, Damásio H, Anderson SW. Insensitivity to future consequences following damage to human prefrontal cortex. Cognition 1994;50:7-15.
26.	Buelow MT, Suhr JA. Construct validity of the Iowa gambling task. Neuropsychol Rev 2009;19:102-14.
27.	Czernecki V, Pillon B, Houeto JL. Does bilateral stimulation of the subthalamic nucleus aggravate apathy in Parkinson’s disease? J Neurol Neurosurg Psychiatry 2005;76:775-9.
28.	Evens R, Stankevich Y, Dshemuchadse M, Storch A, Wolz M, Reichmann H, et al. The impact of Parkinson’s disease and subthalamic deep brain stimulation on reward processing. Neuropsychologia 2015;75:11-9. doi: 10.1016/j.neuropsychologia.2015.05.005.
29.	Oyama G, Shimo Y, Natori S, Nakajima M, Ishii H, Arai H, et al. Acute effects of bilateral subthalamic stimulation on decision-making in Parkinson’s disease. Parkinsonism Relat Disord 2011;17:189-93.
30.	Atkinson-Clement C, Cavazzini É, Zénon A, Witjas T, Fluchère F, Azulay JP, et al. Effects of subthalamic nucleus stimulation and levodopa on decision-making in Parkinson’s disease. Mov Disord 2019;34:377-85. doi: 10.1002/mds.27625.
31.	Boller JK, Barbe MT, Pauls KAM, Reck C, Brand M, Maier F, et al. Decision-making under risk is improved by both dopaminergic medication and subthalamic stimulation in Parkinson’s disease. Exp Neurol 2014;254:70-7. doi: 10.1016/j.expneurol.2014.01.005
32.	Buelow MT, Frakey LL, Grace J, Friedman JH. The contribution of apathy and increased learning trials to risky decision-making in Parkinson’s disease. Arch Clin Neuropsychol 2014;29:100-9.
33.	Irmen F, Horn A, Meder D, Neumann WJ, Plettig P, Schneider GH, et al. Sensorimotor subthalamic stimulation restores risk-reward trade-off in Parkinson’s disease. Mov Disord 2019;34:366-76.
34.	Sallum I, Mata F, Malloy-Diniz L, Miranda D. Staying and shifting patterns across IGT trials distinguish children with externalizing disorders from controls. Front Psychol 2013;4:899.