Deep brain stimulation (DBS) at the subthalamic nucleus (STN) improves motor function and the quality of life of patients with advanced Parkinson’s disease (PD) [
The decline in VF observed in PD patients who undergo DBS is not well understood. Studies have hypothesized that this impairment is due to a possible lesion effect from surgery and/or an effect of the neurostimulator parameters, for instance, the frequency of stimulation [
Low-frequency stimulation has been associated with improved motor symptoms, including freezing of gait and swallowing, in patients with STN-DBS [
VF tasks are often used as operating measures of language and executive functions [
The various VF tasks may provide different types of information regarding cognition because each VF task requires accessing specific lexical and/or semantic representations according to the criteria. The VF tasks activate overlapping areas of the frontal brain regions, but different word retrieval criteria likely activate additional distinct regions [
This study aimed to analyze the impact of low-frequency (60 Hz) and high-frequency (130 Hz) STN-DBS on VF tasks of PD patients. We hypothesized that PD patients would present VF deficits due to the frontosubcortical impairment caused by the disease. Thus, if there is an influence of the frequency of stimulation in our population, the frequency should affect the VF tasks that rely more on frontosubcortical functions, as required by the phonemic and action VFs. To test this hypothesis, not only is it important to assess phonemic and action VF tasks, but also the VF tasks that we hypothesized would not be affected by the frequency of stimulation, that is, semantic and unconstrained VFs.
The present study was a randomized double-blinded experimental study. The study was conducted with outpatients from the Neurology Service of
The study included 20 patients with idiopathic PD diagnosed according to the criteria of the UK Parkinson’s Disease Society Brain Bank [
Exclusion criteria included the abuse of illicit drugs or benzodiazepines within the last six months, the presence of auditory impairment, as evaluated by an audiometric screening performed by an audiologist, the presence of visual impairment, a clinical diagnosis of depression or the presence of important signs or symptoms of depression (measured according to the 17-item Hamilton Depression Rating Scale, with a cutoff of 23 for very severe depression) [
This study was conducted in a randomized, double-blinded manner. The order of the initial DBS conditions was defined by a medical student using the website
After adjusting the frequency, the participants waited one hour to carry out the VF tasks. They then performed the following VF tasks: phonemic VF (FAS version and letter P version), semantic VF (animals), unconstrained VF, and action fluency. For the FAS version of the phonemic VF task, the participants were asked to say words beginning with the letters “F,” “A,” and “S” for one minute for each letter. The final score was the total number of words beginning with “F,” “A,” or “S” that the participants were able to say [
Demographic (sex, age, and education), cognitive (MMSE), and clinical (time of disease in years, time after surgery in months, Hamilton Depression Rating Scale, and the levodopa-equivalent dose (LED)) variables were considered in secondary analyses. The LED was measured as mg/day and was calculated using conversion formulae [
The ethics committees of out institution approved this study, and all participants gave written informed consent.
Statistical analyses were performed using the Statistical Package for Social Sciences (SPSS version 21.0) with a significance level of 5% (
The initial study sample consisted of 25 individuals; however, 5 were excluded for not meeting the inclusion criteria. Three included patients were not able to complete some of the verbal fluency tasks in both frequency conditions. However, these participants were still included in the data analysis. All participants were on levodopa drugs, only 6 participants were on amantadine (mean dose of 283.33 mg/day), and no one was on anticholinergics or antipsychotics. The baseline characteristics of the participants are presented in Table
Baseline characteristics of the participants.
Variables | Mean | Range |
---|---|---|
Sex, male | 16 (80) | — |
Age | 56.65 | 31–75 |
Education | 10.10 | 2–22 |
Time of disease, years | 15.30 | 10–29 |
Levodopa equivalent dose, mg/day | 1165.00 | 300–2300 |
Time after surgery, months | 2.21 | 0–7 |
MMSE | 26.45 | 21–30 |
Hamilton Depression Rating Scale | ||
Normal | 14 (70) | — |
Mild depression | 4 (20) | — |
Moderate depression | 2 (10) | — |
Amplitude (V), left | 3.02 (0.65) | 1–3.90 |
Amplitude (V), right | 2.98 (0.60) | 1.80–3.60 |
Pulse width ( | 79.50 (17.61) | 60–120 |
Pulse width ( | 81.00 (17.14) | 60–120 |
Frequency (Hz), left and right | 124.00 (26.04) | 60–180 |
SD: standard deviation; MMSE: Mini-Mental State Examination.
Stimulation parameters at the moment of inclusion and verbal fluency outcomes of each subject.
Subject | Sex | Age | Frequency (Hz) | Stimulation contacts (all cases +) | Amplitude (V) | Pulse width ( | Verbal fluency outcomes (at 60 Hz) | |||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
L | R | L | R | L | R | Action | Phonemic (P) | Phonemic (FAS) | ||||
1 | M | 49 | 90 | 3- | 11- | 3.0 | 3.0 | 60 | 60 | — | | |
2 | F | 66 | 110 | 3- | 10- | 3.6 | 1.8 | 90 | 90 | | | |
3 | M | 75 | 130 | 0-, 1- | 8-, 9- | 3.3 | 3.3 | 60 | 60 | ND | | ND |
4 | M | 65 | 130 | 2-, 3- | 11- | 3.0 | 3.6 | 60 | 90 | | | |
5 | F | 62 | 110 | 2- | 11- | 2.8 | 2.6 | 60 | 60 | ND | | ND |
6 | M | 53 | 180 | 1- | 9-, 10- | 2.6 | 2.7 | 90 | 90 | | | |
7 | M | 31 | 140 | 2-, 3- | 8-, 9- | 3.6 | 3.6 | 90 | 120 | | | |
8 | M | 50 | 60 | 0- | 8-, 9- | 2.5 | 3.0 | 90 | 90 | | | |
9 | M | 70 | 130 | 1-, 2- | 9-, 10- | 3.0 | 3.2 | 90 | 90 | | ND | |
10 | M | 47 | 110 | 0- | 11- | 2.8 | 2.8 | 60 | 60 | | | |
11 | M | 46 | 140 | 0- | 9- | 2.6 | 3.0 | 90 | 90 | | | |
12 | M | 70 | 110 | 3- | 11- | 3.6 | 3.6 | 90 | 90 | ND | | |
13 | M | 59 | 160 | 2- | 8- | 3.5 | 3.5 | 120 | 90 | | | ND |
14 | M | 47 | 110 | 1- | 9- | 1.0 | 2.0 | 60 | 90 | — | — | — |
15 | M | 61 | 120 | 2- | 11- | 3.4 | 3.4 | 60 | 60 | | | |
16 | M | 66 | 160 | 2- | 9- | 3.2 | 3.6 | 90 | 90 | | ND | ND |
17 | F | 59 | 120 | 2- | 5- | 3.2 | 3.2 | 90 | 90 | | | |
18 | F | 48 | 130 | 3- | 8- | 3.9 | 2.2 | 60 | 60 | — | — | — |
19 | M | 51 | 110 | 2- | 10- | 3.6 | 3.6 | 90 | 90 | | | |
20 | M | 58 | 130 | 0- | 8- | 2.2 | 2.0 | 90 | 60 | — | | |
F: female; L: left; M: male; R: right;
The first and second sets of VF tasks were compared to determine if there was a practice/learning effect due to the repetition of the tasks. The results showed that there was no significant difference between the two sets of tasks for any of the VF tasks (Table
Comparisons of verbal fluency tasks between moments of administration.
Verbal fluency task | Moment 1 | Moment 2 | Difference | CI 95% | |
---|---|---|---|---|---|
Mean ± SD | Mean ± SD | ||||
Phonemic, P | 14.53 ± 1.61 | 14.47 ± 1.89 | 0.05 | −2.96–3.07 | 0.973 |
Phonemic, FAS | 25.26 ± 2.88 | 25.39 ± 15.33 | −0.11 | −3.36–3.15 | 0.949 |
Semantic, animals | 13.26 ± 1.07 | 13.42 ± 1.22 | −0.16 | −1.86–1.54 | 0.856 |
Unconstrained | 29.63 ± 2.50 | 27.58 ± 3.00 | 2.05 | −1.76–5.87 | 0.292 |
Action | 8.39 ± 1.00 | 9.11 ± 1.30 | −0.72 | −2.65–1.21 | 0.463 |
SD: standard deviation; CI: confidence interval.
Table
Comparisons of the verbal fluency tasks and UPDRS-III between the different frequencies of SNT-DBS.
Variables | 60 Hz | 130 Hz | Difference | CI 95% | |
---|---|---|---|---|---|
Mean ± SE | Mean ± SE | ||||
UPDRS III, total | 34.33 ± 4.74 | 35.44 ± 4.30 | −1.11 | −9.38–7.15 | 0.792 |
UPDRS III, tremor | 2.72 ± 1.20 | 2.00 ± 1.09 | 0.72 | −2.19–3.63 | 0.627 |
UPDRS III, gait | 1.28 ± 0.26 | 1.61 ± 0.33 | −0.33 | −0.80–0.13 | 0.157 |
UPDRS III, pull test | 1.28 ± 0.29 | 1.83 ± 0.34 | −0.56 | −1.12–0.00 | 0.052 |
Phonemic VF, P | 16.53 ± 1.82 | 12.47 ± 1.56 | 4.05 | 1.65–6.45 | |
Phonemic VF, FAS | 26.84 ± 3.36 | 23.79 ± 2.79 | 3.05 | 0.10–6.00 | |
Semantic VF, animals | 13.70 ± 1.20 | 12.89 ± 1.08 | 0.89 | −0.76–2.55 | 0.290 |
Unconstrained VF | 29.63 ± 2.92 | 27.58 ± 2.59 | 1.95 | −1.76–5.87 | 0.292 |
Action VF | 9.94 ± 1.22 | 7.56 ± 1.94 | 2.39 | 0.77–4.00 | |
CI: confidence interval; SE: standard error; UPDRS: Unified Parkinson’s Disease Rating Scale; VF: verbal fluency;
The performances for the VF tasks according to the STN-DBS frequency are shown in Figure
Distribution of delta values for phonemic and action fluency tasks.
Phonemic VF | Phonemic VF | Action VF | |
---|---|---|---|
P version | FAS version | ||
Delta value, mean | 4.05 (5.50) | 3.05 (6.73) | 2.40 (3.60) |
Delta classification, | |||
Improvement at 60 Hz | 14 (70) | 11 (55) | 10 (50) |
Worsening at 60 Hz | 2 (10) | 3 (15) | 3 (15) |
No difference | 2 (10) | 4 (20) | 3 (15) |
SD: standard deviation; VF: verbal fluency.
Patients’ performances on VF tasks at low and high stimulation frequencies. (a) Performance of the entire sample for each VF task.
In the correlational analysis, we included the delta values of the VF task outcomes that showed significant differences between the frequency conditions (P and FAS versions of the phonemic fluency tasks and the action fluency task) to determine if the difference was correlated with additional variables (including the delta values of UPDRS-III). The delta value of the FAS version of the phonemic VF task was negatively associated with age (
Correlational analysis between VF tasks (phonemic and action), demographic, cognitive, and clinical measures.
Phonemic - P | Phonemic - FAS | Action | ||||
---|---|---|---|---|---|---|
| | | | | | |
Age | −0.207 | 0.382 | −0.473 | | −0.352 | 0.152 |
Education | −0.192 | 0.416 | 0.300 | 0.212 | −0.021 | 0.935 |
MMSE | −0.014 | 0.953 | 0.215 | 0.377 | 0.108 | 0.669 |
Time of disease, years | 0.400 | 0.081 | 0.358 | 0.132 | −0.036 | 0.887 |
Time after surgery, months | 0.288 | 0.218 | 0.440 | 0.060 | 0.284 | 0.253 |
Levodopa equivalent dose | 0.351 | 0.140 | 0.131 | 0.592 | −0.003 | 0.992 |
HDRS, total score | 0.011 | 0.964 | 0.078 | 0.757 | −0.138 | 0.599 |
UPDRS, total | −0.686 | | −0.342 | 0.165 | −0.058 | 0.825 |
UPDRS, tremor | −0.200 | 0.426 | −0.133 | 0.600 | 0.020 | 0.939 |
UPDRS, gait | −0.378 | 0.122 | 0.004 | 0.986 | 0.003 | 0.990 |
UPDRS, pull test | −0.312 | 0.207 | −0.479 | 0.068 | −0.321 | 0.209 |
HDRS: Hamilton Depression Rating Scale; MMSE: Mini-Mental State Examination; UPDRS: Unified Parkinson’s Disease Rating Scale;
The present study aimed to investigate the impact of modulating the frequency of STN-DBS on the performance of VF tasks in patients with PD. We assessed the effect of low-frequency stimulation of 60 Hz compared to high-frequency stimulation of 130 Hz in patients who had undergone bilateral STN-DBS in the medication-on state. We found that low-frequency stimulation had a positive impact on phonemic and action fluency, and this effect was not due to practice. Furthermore, we observed different outcome patterns for the VF tasks based on the frequency conditions, which could not be explained by the demographic, cognitive, and clinical variables that were studied here.
Previous studies have pointed to a decline in VF after STN-DBS surgery in PD patients, although the reason behind this decline is not well understood. There are many methodological differences among such studies, such as evaluations performed while stimulation is “on” or “off,” at pre- and postsurgical time points, and with or without a control group [
Greater declines in VF over time in STN-DBS patients compared to nonsurgical PD patients have been reported, which suggests that VF impairment is related to the DBS intervention [
The frequency of stimulation for treatment of PD has been studied in other clinical situations regarding STN-DBS. For example, 60 Hz stimulation, compared with the routine 130 Hz, improved swallowing function and freezing of gait in patients with PD who underwent bilateral STN-DBS [
When we compared the VF scores and the motor performances between the low-frequency (60 Hz) and high-frequency (130 Hz) stimulation trials, we found that phonemic and action fluency significantly declined for 130 Hz stimulation. However, no significant difference was found between stimulation frequencies for the semantic and unconstrained VF tasks. Our findings indicate that the influence of the stimulation frequency relies more heavily on specific frontosubcortical pathways involved in lexical-word and action-semantic processes, as there was an influence of frequency on phonemic and action VF tasks but not on semantic and unconstrained VF tasks. As we expected, the frequency affected the VF tasks that rely more heavily on frontosubcortical functions, which are impaired in PD, supporting our
Phonemic and action fluency, which were hampered by high-frequency neurostimulation, are both tasks that involve frontal circuits and that rely more heavily on executive functions. Semantic VF depends on lexical-semantic processes and temporal circuits [
In the present study, no significant difference was found in the motor performance of the patients in relation to the stimulation frequency. One hypothesis for this finding is that the patients were under the effect of dopaminergic medication during the evaluation. If there was no effect of the medication, the low-frequency stimulation (60 Hz) would be expected to correspond to worse motor symptoms, whereas an improvement in motor performance for the high-frequency stimulation would be expected [
The comparison analyses revealed that most patients performed better in phonemic and action VF tasks for the low-frequency condition. However, a few of the participants presented an opposite pattern (worse VF scores for low-frequency stimulation) or no difference between the conditions. In an attempt to elucidate the cause of these different patterns, correlational analyses were conducted with the delta values of the phonemic and action VF scores. The scores from the FAS version of the phonemic VF task were negatively correlated with age, indicating a possible effect of age on the benefit of low-frequency stimulation; that is, older participants may exhibit lower improvements in performance for low-frequency stimulation compared to younger patients. The scores from the P version of the phonemic VF task were negatively correlated with UPDRS-III total score, indicating that increased improvements in motor performance were associated with smaller improvements in this phonemic VF task for low-frequency stimulation. This latter finding indicates that motor performance and phonemic VF scores (P version) are characterized by opposite outcomes at low-frequency stimulation. Future studies should seek identifying the factors that explain the different VF improvement profiles for varying frequency stimulation conditions by studying larger samples of PD patients with STN-DBS.
Our study is the first one that called attention to different outcomes of verbal fluency when frequency conditions were compared. Many aspects may influence the modulation of frequency on verbal fluency. Because the STN is thought to have separate functional subregions [
Our results should be interpreted in light of some limitations. First, the improvement in the VF task for the low-frequency condition could be due to an improvement in other cognitive functions, such as attention. However, this study did not evaluate other cognitive functions. We chose to utilize a less-extensive assessment because some patients do not tolerate adjustments in stimulation frequency for long periods of time. Second, the participants were not evaluated in the DBS-off condition so that they were not exposed to unpleasant symptoms for long periods of time, and this study did not include a control group. The lack of information for the DBS-off condition and the lack of a control group do not allow surgical effects to be assessed. Third, the administration order of the VF tasks was the same in both conditions. There is the possibility of an order effect, but based on a previous study, we do not believe an order effect occurred, at least pertaining to the action VF task [
In summary, the results of the present study led to two important conclusions. First, the frequency of STN-DBS affects phonemic and action fluency in PD patients. Second, low-frequency (60 Hz) stimulation is associated with less negative side effects on VF than high-frequency (130 Hz) stimulation. Therefore, whenever possible, low-frequency stimulation should be the first choice for PD patients, especially for patients who present any cognitive impairments, such as reduced VF, in their daily activities. Future studies utilizing larger sample populations and those incorporating longer study periods should be performed to investigate stimulation effects on VF with regard to electrode position in the STN and other stimulation parameters (amplitude and pulse width).
The authors declare that they have no competing interests.
The authors would like to thank the patients for their participation, the