Mallet et al. 10.1073/pnas.0610849104.

Supporting Information

Files in this Data Supplement:

SI Materials and Methods

Patient 1. Rest tremor of the left leg started at age 47 in this patient, a project manager without a previous medical history. His response to levodopa treatment (1,225 mg/day) was excellent (UPDRS III: off, 36; on, 4) but motor fluctuations and dyskinesias developed within 2 years. Neurosurgery for bilateral STN stimulation was performed 8 years after onset of the disease. A preoperative psychiatric evaluation based on the semi-structured Mini-International Neuropsychiatric Interview (MINI 5.0.0.) and a non-structured interview did not reveal any personal or familial psychiatric history, and in particular no bipolar disorder. After surgery, bilateral stimulation through the ventral electrodes markedly improved his motor disability (UPDRS III = 15), but, within several days, the patient became irritable, talkative, impulsive, unable to concentrate and insomniac. He made a lot of jokes and irrelevant remarks, but had no insight into the pathological nature of his loss of social inhibition (Patient: "I didn't sleep much last night, we lived it up with the neighbors ... this morning I woke up feeling great (...)". Spouse: "He's very cheerful, like somebody who's euphoric, a little too much, but he gets irritated and loses his temper very fast, he boils over then it's finished"). Hypomania was not ameliorated by valproic acid (1,000 mg/day). When stimulation was switched from the ventral to the dorsal contact, the patient calmed down within 15 minutes and his logorrhea subsided within a few hours, but his motor improvement was not significantly affected (Spouse: "He's better balanced, calmer, ... no longer aggressive, it's wonderful"). At the time of the experiment, six months after surgery, valproic acid was discontinued, parkinsonian motor disability was improved by 78% (parameters left/right electrode: 2.1/2.7 V, 60/60 µs, 130/130 Hz; daily levodopa equivalent dose: 150 mg, i.e. levodopa 50 mg/carbidopa 12.5 mg three times daily at 7 a.m. 12 a.m. 7 p.m.) with mild dyskinesias. The patient was working full time and had a normal family life. His cognitive status was considered normal (Mini-Mental State Examination (MMSE): 30/30; Mattis : 144/144; frontal score: 46/50). The treatment was strictly as usual the day of PET experiment. The patient took no additional medication, in particular no benzodiazepines. Only the contacts of the left electrode were switched in this patient during the PET protocol, because a change to the ventral contact in the right hemisphere severely worsened parkinsonian motor disability.

Patient 2. Parkinson's Disease (PD) began in this 55-year-old professional musician at age 37 with bradykinesia on the right side. Levodopa treatment (begun at age 41) allowed the patient to continue to perform in public. His response to levodopa treatment (1,100 mg/day, UPDRS III = off 37, on 4), was excellent. Because of severe motor fluctuations and dyskinesias, surgery for bilateral STN stimulation was performed when the patient was 55. A preoperative psychiatric evaluation based on a semi-structured interview (MINI 5.0.0.) and a non structured interview did not reveal any familial psychiatric history and no personal bipolar history. He had an episode of severe depression at age 43,6 years after the onset of PD motor symptoms, without psychotic features, that was successfully treated by fluoxetine 20mg/d. He was still treated at the time of surgery. Stimulation was started bilaterally through the ventral electrodes. A few hours later, the patient developed persistent hypomania. The patient experienced this as a pleasurable state, felt more creative, and was full of ideas for musical compositions. The excitation subsided within a few hours after stimulation was switched to the dorsal electrodes. Six months after surgery, motor disability improved 86% with stimulation alone (parameters left/right electrode: 2.5/2.5 V, 60/60 µs, 130/130 Hz) and 100% after medication was added (daily levodopa equivalent dose: 300 mg, i.e. levodopa 50 mg/carbidopa 12.5 mg four times daily and pergolide 0.25 mg four times daily at 8 a.m. 12 a.m. 4 p.m. 8 p.m.; fluoxetine: 20 mg 8 a.m.). The treatment was strictly as usual the day of PET experiment. The patient took not additional medication, in particular no benzodiazepines. At the time of the experiment, nine months after surgery, social behavior, cognition (MMSE: 30/30; Mattis: 143/144; frontal score: 50/50) and mood were normal, and the patient still performed professionally at a high level.

Diagnosis of hypomania. Stimulation through the ventral contact provoked very similar reactions in both of the patients studied here who otherwise had no personal or familial antecedents of bipolar disorder according the preoperative evaluation based on a semi-structured psychiatric interview (MINI 5.0.0.) , and no immediate psychotropic or behavioral effects were noted or self-perceived after stimulation was switched from the dorsal to the ventral contact. However, a few hours later, patients became more talkative and developed a persistent abnormal elevated or irritable mood that lasted for several weeks during the postoperative period, and therefore met Criterion A (at least 4 days) for a hypomanic episode according to the DSM-IV . Additional clinical features differed somewhat between the two patients. Patient 1 became slightly irritable, made jokes, played on words, made irrelevant remarks and couldn't concentrate. Patient 2 increased his involvement in goal-directed activities and experienced stimulation through the ventral contact as pleasurable; he felt more creative and was full of ideas for musical compositions while this contact was activated. These features, associated with a decreased need for sleep and pressure to keep talking in both patients, met DSM-IV Criterion B. Their mood and the change in the way they functioned during this period were clearly unusual (Criterion C). The changes in mood and functioning were observed by the spouse of patient 1 and the close entourage of patient 2 (Criterion D). The episodes were not severe enough in either patient to require hospitalization, and there were no psychotic features (Criterion E). A few hours after stimulation was switched back to dorsal contact, both patients had returned to their baseline psychiatric and motor condition. According to DSM-IV criterion F, the hypomanic features reported in our two patients should be strictly distinguished from a hypomanic episode, since they were evoked by stimulation of the subthalamic nucleus, judged to be etiologically related to the mood disturbance. Therefore, these episodes have been diagnosed as deep brain stimulation-induced mood disorders with manic features, according to DSM-IV criteria, in the absence of a concomitant change in medication.

Unified Parkinson Disease Rating Scale part III (UPDRS III) . Part III of the Unified Parkinson Disease Rating Scale (UPDRS III) is the most accepted and widely used clinical rating scale for the severity of parkinsonian signs . It can be performed in less than 10 minutes, has been extensively validated and comprises 14 items. Seven items evaluate midline symptoms (speech, facial expression, ability to get up from a chair, posture, gait, equilibrium), two items rate rest, postural and action tremor, one item rates rigidity, four items rate repetitive movements on both sides, and one item rates overall bradykinesia. Rest tremor and rigidity are rated separately for each limb and for the neck/head. The score ranges from 0 to 108; the higher the score the more severe the disease.

Young Mania Rating Scale (YMRS) . The YMRS evaluates the severity of manic symptoms and the effect of treatment. Severity is determined by a personal interview that takes 15-30 minutes, depending on the patient's handicap (here the severity of symptoms during the hours preceding the interview), and is based on a checklist of 11 items. Seven are rated 0-4 (elevated mood, increased motor activity and energy, sexual interest, sleep, language and thought disorder, appearance, insight) and four 0-8 (irritability, speech, content, disruptive-aggressive behavior); the higher the score, the greater the severity (maximun score, 60).

Neuropsychological assessment of changes related to the hypomanic state by the continuous performance test (CPT-II). The neuropsychological status of the patients was evaluated with Conners' continuous performance test (CPT-II) of attention . The CPT-II evaluates cognitive functions that are impaired during mania and has shown to be sensitive to changes in patients with manic features . The CPT-II takes only 14 minutes to complete and can be administered without limitations on the number of repetitions and could thus be used at the time of PET experiment could be to the experimental procedures. It consists of successive discrimination tasks in which patients are required to press the space bar on a computer keyboard when any letter except the target letter 'X' appears. The inter-stimulus intervals (ISIs) are 1, 2 and 4 seconds with a display time of 250 milliseconds. There are 6 blocks, with 3 sub-blocks, each containing 20 trials (letter presentations). The order in which the different ISIs are presented varies among blocks. Speed, errors and the consistency of responses are measured, as follows: (i) Omissions result from the failure to respond to target letters (i.e. non-Xs). (ii) Commission errors are made when responses are given to non-targets (i.e. Xs). (iii) Overall Hit Reaction Time (RT) is the average speed of correct responses for the entire test. (iv) Variability is a measure of response speed consistency. The higher the variability, the greater the inconsistency in the response speed. (v) Hit RT Block Change measures change in reaction time across the duration of the test. High values indicate a substantial slowing in reaction times. Low values indicate that responses were more rapid as the test progressed. (vi) Hit RT Standard Error by Block detects change in response consistency over the duration of the test. High values indicate a substantial loss of consistency as the test progressed. Low values on this measure indicate sustained or improved response consistency. Any reaction time that is less than 100 ms constitutes a perseverative response. Given normal physiology, a perseverative response usually reflects either anticipation or random choice. The CPT-II software automatically performs a discriminant function analysis to provide an overall index of whether the profile of the patient best fits a pathological or non- pathological pattern. The overall index permits classification: values below 8 generally indicate no problems with attention, values between 8 and 11 indicate that there may be problems, and values above 11 suggest the presence of attention deficits.

Mattis Dementia Rating Scale . This scale is a standardized, structured mental status examination that provides a global measure of dementia. It takes 30 minutes for a patient without dementia. The score (0-144) is the sum of correct responses on tests of five cognitive functions: attention, initiation and perseveration, construction, conceptualization, and memory.

Mini-Mental State Examination (MMSE) . The MMSE is a very brief , easily administered mental status examination that is widely used in clinical practice. It is a structured scale that consists of 30 items covering seven categories: orientation in space (state, country, town, hospital, and floor) and time (year, season, month, day, and date), memorisation(repeat three words), attention and concentration (serial substraction of 7 from 100, or spelling the word world backward), recall ( the previously repeated words), language (name two objects, repeat a phrase, read loud and understand a sentence, write a sentence, and follow a three-step order) and visual construction.

Frontal Score . The frontal score evaluates executive functions: concept or set formation (Wisconsin card sorting test), category fluency (number of words in a minute; names of fruits), verbal fluency (words beginning with V), graphic and motor function.

PET Data Analysis. PET data were analyzed by statistical parametric mapping (SPM2, Wellcome Department of Imaging Neuroscience, 2003; http://www.fil.ion.ucl.ac.uk/spm/software/spm2/) by Matlab 6.5 software (Mathworks, USA) on a Linux (Red Hat 9) workstation. Individual scans were automatically realigned to correct for motion artefacts, and were normalized spatially to the standard PET template provided in SPM with an affine transform. Images were then smoothed with a 12 mm full width at half maximum isotropic three-dimensional Gaussian filter and resampled to 2 mm ´2 mm ´ 2 mm voxels. Variations in the global concentration of radiotracer between scans were corrected using proportional scaling (nominal value: 50 ml/100 g/min). The differences between relative regional cerebral blood flow (rCBF) in the two experimental conditions (hypomanic, H, vs. normal, N) were tested statistically for each voxel using a fixed-effects model treating subject as a confound covariate. There were 6 replications per subject and condition leaving 21 degrees of freedom. Proportional threshold masking was applied with a criterion set to 80% of the mean global value and replicated measures were corrected for non-sphericity. SPM{t}maps across the two patients were computed for each contrast (H>N and N>H). The voxelwise significance threshold was set at P < 0.00001 (T=5.47; uncorrected for multiple comparisons) for a minimum cluster of 100 voxels. Final results were presented in the standard Montreal Neurological Institute (MNI) coordinate space, which is parallel to Talairach space, , and were labelled with the Automatic Anatomical Labeling toolbox (AAL v1.0, ). Brodmann Area labeling was performed with MRIcro software (v1.39 ). The exact locations of sub-cortical activations were determined for each patient by fitting images from the three dimensional basal ganglia atlas described above. The PET data acquired in the hypomanic and non-manic states were contrasted using the same height and spatial thresholds. The resulting individual SPM{t} maps were merged with the anatomical MRI, which had previously been coregistred with the PET acquisition geometry, and were identified according to the atlas. Brain activity was designated as increasing or decreasing according to the sign of the contrast used in the statistical parametric mapping .