See Table S1 for detailed clinical and demographical data. Patients were hospitalized either for a biopsy, in order to determine the nature of the tumor (n = 5), or for the surgical ablation of buy IWR-1 the tumor (n = 18). Tumors were gliomas in all patients but one, in whom the tumor was metastatic. The precise glioma types were (grades are given following the World Health Organization

classification): 12 oligoastrocytomas (grade 2: n = 7; grade 3: n = 5), three oligodendrogliomas (all grade 2), two astrocytomas (grade 2 for one and grade 3 for the other), one pilocytic astrocytoma (grade 1), and two glioblastomas (grade 4). For two gliomas, the precise type could not be determined (one was grade 2 and the other grade 3). A majority of patients (15/23) were under preventive antiepileptic medication because of a history of tumor-related seizure. No patient was taking any medication interfering with the dopaminergic system, such as neuroleptics. Patients were tested 29 ± 13 (mean ± SEM) months after the onset of clinical symptoms and 24 ± 12 months after the MRI or computed tomography scan that had confirmed the diagnosis of tumoral mass present in the brain. Patients were AT13387 cost split according to whether the lesion overlapped with the insula (INS group: n = 14) or not (LES group: n = 9). Tumor etiology was globally matched between the two groups, with similar grades (INS: 2.4 ± 0.2; LES: 2.1 ± 0.2; p > 0.3, t test)

and a similar about proportion of oligoastrocytomas (INS: 8/14; LES: 4/9; p > 0.4, chi2-test). We also checked that lesion sizes were comparable between the two groups (INS: 76.6 ± 10.8; LES: 92.0 ± 22.0; p > 0.5, t test). A cohort of healthy subjects was also included (CON group; n = 20). These subjects were matched to INS patients in age (CON: 43.6 ± 2.8; INS: 46.7 ± 3.9; p > 0.5, t test), gender (CON: 12/8; INS: 9/5; p > 0.7, chi2-test), and handedness (CON: 16/4; INS: 11/3; p > 0.9, chi2-test). There was no cognitive impairment in the INS group, as indicated by the normal Mini-Mental State (MMS) score (29.3 ± 0.6). INS patients were not depressed (Hospital Anxiety and Depression

[HAD] depression score: 4.9 ± 0.7), but moderately anxious (HAD anxiety score: 8.2 ± 1.2). Unfortunately, the MMS and HAD scores were only collected for a minority of LES patients (4/9), in whom they were similar to those obtained in the INS group (MMS: 30.0 ± 0.0; HAD depression: 4.5 ± 2.4; HAD anxiety: 10.3 ± 2.9). All lesioned patients but one had a high-definition three-dimensional anatomical T1 MRI scan and a fluid attenuated inversion recovery T2 MRI scan. The scans were acquired on average 39.6 ± 23.6 days before the experiment. Based on both T1 and T2 scans, the tumoral masses were manually segmented on the native anatomical space using MRIcro (http://www.cabiatl.com). The T1 scans were normalized to an anatomical template with the Statistical Parametric Mapping software (SPM8: http://www.fil.ion.ucl.ac.