Article
Reference
Vocal emotion decoding in the subthalamic nucleus: An intracranial ERP study in Parkinson's disease
PERON, Julie Anne, et al.
Abstract
Using intracranial local field potential (LFP) recordings in patients with Parkinson's disease (PD) undergoing deep brain stimulation (DBS), we explored the electrophysiological activity of the subthalamic nucleus (STN) in response to emotional stimuli in the auditory modality.
Previous studies focused on the influence of visual stimuli. To this end, we recorded LFPs within the STN in response to angry, happy, and neutral prosodies in 13 patients with PD who had just undergone implantation of DBS electrodes. We observed specific modulation of the right STN in response to anger and happiness, as opposed to neutral prosody, occurring at around 200–300 ms post-onset, and later at around 850–950 ms post-onset for anger and at around 3250–3350 ms post-onset for happiness. Taken together with previous reports of modulated STN activity in response to emotional visual stimuli, the present results appear to confirm that the STN is involved in emotion processing irrespective of stimulus valence and sensory modality.
PERON, Julie Anne, et al . Vocal emotion decoding in the subthalamic nucleus: An intracranial ERP study in Parkinson's disease. Brain and Language , 2017, vol. 168
DOI : 10.1016/j.bandl.2016.12.003
Available at:
http://archive-ouverte.unige.ch/unige:95989
Disclaimer: layout of this document may differ from the published version.
1 / 1
Vocal emotion decoding in the subthalamic nucleus: An intracranial ERP study in Parkinson’s disease
Julie Péron
a,b,⇑, Olivier Renaud
c, Claire Haegelen
d,e, Lucas Tamarit
b, Valérie Milesi
a,b,
Jean-François Houvenaghel
f,g, Thibaut Dondaine
f,g,h, Marc Vérin
f,g, Paul Sauleau
f,i,1, Didier Grandjean
a,b,1a’Neuroscience of Emotion and Affective Dynamics’ Laboratory, Department of Psychology & Swiss Center for Affective Sciences, University of Geneva, 40 bd du Pont d’Arve, 1205 Geneva, Switzerland
bNeuropsychology Unit, Department of Neurology, University Hospitals of Geneva, Rue Gabrielle-Perret-Gentil 4, 1205 Geneva, Switzerland
cMethodology and Data Analysis Unit, Department of Psychology, University of Geneva, 40 bd du Pont d’Arve, 1205 Geneva, Switzerland
dNeurosurgery Department, Pontchaillou Hospital, Rennes University Hospital, rue Henri Le Guilloux, 35033 Rennes, France
eINSERM, LTSI U1099, Faculty of Medicine, CS 34317, University of Rennes I, F-35042 Rennes, France
f’Behavior and Basal Ganglia’ Research Unit (EA 4712), University of Rennes 1, Rennes University Hospital, rue Henri Le Guilloux, 35033 Rennes, France
gNeurology Department, Pontchaillou Hospital, Rennes University Hospital, rue Henri Le Guilloux, 35033 Rennes, France
hAdult Psychiatry Department, Guillaume Régnier Hospital, 108 avenue du Général Leclerc, 35703 Rennes, France
iPhysiology Department, Pontchaillou Hospital, Rennes University Hospital, rue Henri Le Guilloux, 35033 Rennes, France
a r t i c l e i n f o
Article history:
Received 27 July 2016 Revised 22 November 2016 Accepted 12 December 2016 Available online 12 January 2017
Keywords:
Deep brain stimulation Emotional prosody Local field potentials Parkinson’s disease Subthalamic nucleus
a b s t r a c t
Using intracranial local field potential (LFP) recordings in patients with Parkinson’s disease (PD) under- going deep brain stimulation (DBS), we explored the electrophysiological activity of the subthalamic nucleus (STN) in response to emotional stimuli in theauditorymodality. Previous studies focused on the influence ofvisualstimuli. To this end, we recorded LFPs within the STN in response to angry, happy, and neutral prosodies in 13 patients with PD who had just undergone implantation of DBS electrodes. We observed specific modulation of therightSTN in response to anger and happiness, as opposed to neutral prosody, occurring at around 200–300 ms post-onset, and later at around 850–950 ms post-onset for anger and at around 3250–3350 ms post-onset for happiness. Taken together with previous reports of modulated STN activity in response to emotional visual stimuli, the present results appear to confirm that the STN is involved in emotion processing irrespective of stimulus valence and sensory modality.
Ó2017 Elsevier Inc. All rights reserved.
1. Introduction
Emotional prosodyis defined as suprasegmental changes in the course of a spoken utterance, encompassing intonation, amplitude, envelope, tempo, rhythm, and voice quality (Grandjean, Banziger,
& Scherer, 2006). The perception and decoding of emotional pro- sody has been studied through functional magnetic resonance imaging (fMRI) and patient studies, allowing researchers to delin- eate a distributed neural network involved in its identification and
recognition (for a recent review, see Witteman, Van Heuven, &
Schiller, 2012). Accordingly, models of emotional prosody process- ing have long postulated the existence of multiple information pro- cessing stages related to different levels of representations (Schirmer & Kotz, 2006; Wildgruber, Ethofer, Grandjean, &
Kreifelts, 2009). Following the processing of auditory information within the primary auditory regions (Bruck, Kreifelts, &
Wildgruber, 2011; Wildgruber et al., 2009), including the activa- tion of predominantly right-hemispheric primary and secondary auditory regions (for a review, seeWitteman, van Ijzendoorn, van de Velde, van Heuven, & Schiller, 2011), two prosody decoding steps have been identified. The first step, related to therepresenta- tion of meaningful suprasegmental acoustic sequences, is thought to involve projections from the superior temporal gyrus to the ante- rior superior temporal sulcus. These cortical structures are thought to represent the so-called temporal voice area that encompasses voice-sensitive neuronal populations (Belin & Zatorre, 2000;
Grandjean et al., 2005). In the second step, emotional information
http://dx.doi.org/10.1016/j.bandl.2016.12.003 0093-934X/Ó2017 Elsevier Inc. All rights reserved.
Abbreviations: AC, anterior commissure; BG, basal ganglia; CT, computed tomography; DBS, deep brain stimulation; ERP, event-related potential;f0, funda- mental frequency; fMRI, functional magnetic resonance imaging; FWER, familywise error rate; IAPS, International Affective Picture System; LFP, local field potential; PC, posterior commissure; PD, Parkinson’s disease; STN, subthalamic nucleus.
⇑ Corresponding author at: Faculty of Psychology and Educational Sciences, 40 bd du Pont d’Arve, 1205 Geneva, Switzerland.
E-mail address:[email protected](J. Péron).
1 These authors contributed equally to this work.
Contents lists available atScienceDirect
Brain & Language
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / b & l
derived at the level of the superior temporal sulcus is made avail- able for higher-order cognitive processes mediated by the right inferior frontal gyrus (Frühholz & Grandjean, 2013) and orbitofron- tal cortex (Ethofer et al., 2006; Grandjean, Sander, Lucas, Scherer, &
Vuilleumier, 2008; Sander et al., 2005; Wildgruber et al., 2004).
This step is thought to involve higher representation such as the explicit evaluation of vocally expressed emotions. In addition to this frontotemporal network, increased activity has been observed within the amygdala in response to emotional prosody (Frühholz, Ceravolo, & Grandjean, 2012; Grandjean et al., 2005; Sander et al., 2005). Furthermore, although they were not the focus of these studies, the involvement of subcortical regions (other than the amygdala) such as the thalamus (Wildgruber et al., 2004) and the basal ganglia (BG) in the processing of emotional prosody has also been reported. The involvement of the caudate and the putamen has repeatedly been observed in healthy participants by using fMRI and in brain-damaged patients by using the electroen- cephalographic method (Bach et al., 2008; Grandjean et al., 2005;
Kotz et al., 2003; Morris, Scott, & Dolan, 1999; Paulmann & Kotz, 2008). More recently, studies exploring the effects of subthalamic nucleus (STN) deep brain stimulation (DBS) in Parkinson’s disease (PD) have highlighted the involvement of the STN in the brain net- work subtending vocal emotions (Bruck, Wildgruber, Kreifelts, Kruger, & Wachter, 2011; Péron, Cekic, et al., 2015; Péron, Frühholz, Vérin, & Grandjean, 2013; Péron et al., 2010).
From convergent evidence in the latter literature,Péron et al.
(2013)have posited that the STN would coordinate neural pat- terns, either synchronizing or desynchronizing the activity of the different neuronal populations responsible for specific emotion components. They claim that the STN plays ‘‘the role of neural rhythm organizer at the cortical and subcortical levels in emotional processing, thus explaining why the BG are sensitive to both the tem- poral and the structural organization of events” (Péron, Frühholz, Ceravolo, & Grandjean, 2015; Péron et al., 2013). This model sug- gests that the BG and, more specifically, the STN, are sensitive to rhythmbecause oftheir intrinsic functional role of rhythm orga- nizer, or coordinator of neural patterns. In a more operational way, this model hypothesizes that (i) the STN is involved in all stages of emotional processing, and (ii) the STN is involved in emo- tion processing irrespective of stimulus valence (positive or nega- tive) and sensory modality (e.g., visual or auditory). To date, these two hypotheses have yet to be tested at the neurophysiological level in the vocal modality.
Studies featuring intracranial recordings of local field potentials (LFPs) in the STN in response to emotional stimuli have investi- gated the impact on the STN’s electrophysiological activity of pro- cessing visualstimuli that induce affective states. These studies have consistently reported event-related desynchronization of alpha activity within the STN between 1000 and 2000 ms post- onset in response to emotional stimuli in the form of both pleasant and unpleasant images drawn from the International Affective Pic- ture System (IAPS) (Brucke et al., 2007; Huebl et al., 2011; Kühn et al., 2005). This effect has been shown to be correlated with strength ofvalence (both positive and negative), but notarousal (Brucke et al., 2007). Furthermore, in the presence of depressed mood, this effect is reduced for positive emotions, but enhanced for negative emotions (Huebl et al., 2011), supporting the idea that the bias toward negative emotions in depression has an electro- physiological signature in this region (Péron et al., 2011). Taken together, these results confirm the STN’s functional role in emotion processing. However, the use ofvisual(emotional) material intro- duces several confounding factors. First, the pictures used in stud- ies exploring the visual modality (see, for example,Benedetti et al., 2004; Kühn et al., 2005) are often not controlled for low-level physical features (brightness, contrast, spatial frequency, etc.). This is methodologically problematic, particularly for the analysis of
neural oscillations in response to emotional stimuli, as these fea- tures are known to influence emotional perception and related brain activities (Delplanque, N’Diaye, Scherer, & Grandjean, 2007). Moreover, the emotional effects reported in the visual domain could be explained either by deficits in visual exploration related to oculomotor abnormalities (Fawcett, Dostrovsky, Lozano, & Hutchison, 2005), or by specific oculomotor patterns generated during the visual exploration of visual emotional stimuli (Van Reekum et al., 2007).
For all these reasons, previous research has left several ques- tions unanswered, and it has yet to be determined whether the STN’s electrophysiological changes in response to emotional stim- uli are modality-specific or supramodal, as suggested by behavioral studies exploring the emotional effects of STN DBS in PD (seePéron et al., 2013for an exhaustive review). To the best of our knowledge, the only study in theauditorymodality was recently performed by Eitan et al. (2013). They reported microelectrode recordings in the STN of 17 PD patients in response to emotional prosody stimuli (onomatopoeias from the Montreal Affective database, Belin, Fillion-Bilodeau, & Gosselin, 2008). The authors reported ventral STN activity in response to emotional versus neutral auditory material, with increases of oscillations observed solely in theright STN; this difference has not been demonstrated in theleft STN.
However, similar to studies of the facial modality, this study did not control for low-level physical features known to have an impact on emotional prosody (e.g., fundamental frequency [f0], energy). Moreover, as mentioned earlier, the authors used single- unit recordings, and to the best of our knowledge, no previous study investigated LFPs in response to emotional auditory stimuli.
The latter type of signal presents the advantage, as compared to single-unit recordings, to reflect the averaged dendrosomatic activity of synaptic signals of large neuronal population (Buzsaki, Anastassiou, & Koch, 2012).
In this context, the aim of the present study was thus to exam- ine the influence of the processing of (both positive and negative) emotional prosody stimuli on event-related potentials (ERPs) in the STN, controlling for low-level physical features crucial for emo- tional prosody decoding. To this end, we explored the electrophys- iological activity (LFPs) of the STN in response to angry, happy, and neutral prosodies, but also to acoustically matched nonhuman syn- thesized stimuli, in 13 PD patients.
The operational hypotheses are mainly based on Péron and col- leagues’ recent model of the STN’s functional role in emotional (prosody) processing(2013), together with previous results from studies featuring intracranial recordings in the STN in response to emotional stimuli (Brucke et al., 2007; Huebl et al., 2011;
Kühn et al., 2005). Péron and colleagues’ model(2013)suggested that the STN would be involved in all stages of emotional process- ing, irrespective of stimulus valence (positive or negative) and sen- sory modality. Accordingly, in the present study we expected to observe STN modulation for both positive and negative emotional prosody stimuli.
Moreover, this model hypothesizes that the STN would partici- pate in the acquisition and expression of sequential, repetitive, motor, and cognitive behaviors, as well as emotional behaviors, eli- cited by external or internal triggers, recruiting and synchronizing the activity of the cortical and subcortical structures required for the relevant process, depending on the nature of the environmen- tal cues and how the individual appraised such characteristics. As a consequence, ‘‘several different[STN]temporal dynamics[are spec- ulated]:(i) sustained, synchronized activity, leading to the creation of a new functional neural pattern; (ii) early, transient synchronized activity, resulting in the activation of an overlearned/innate pattern;
and (iii) later, transient synchronized activity as part of a mechanism that inhibits the overlearned pattern (involving the prefrontal regions)” (Péron et al., 2013, p. 370). Given that we do not explore
the creation of a new functional pattern in this study, nor its inhi- bition, our present study explored the processes related to the acti- vation of an overlearned/innate pattern. Moreover, the temporal dynamics depends on the specific process in action; for example, the activation of a pattern specific to the human voice is related to the coordination of specific brain areas and should occur at a dif- ferent time, probably earlier through ascending auditory pathways (e.g.,Escera, Leung, & Grimm, 2014), than the brain activation of a pattern specific to emotion through, for example, amygdala pre- frontal coupling (e.g.,Courtin, Karalis, Gonzalez-Campo, Wurtz, &
Herry, 2014). As a consequence, we predicted two different tempo- ral dynamics involving different brain patterns: (i) early stimulus- driven activity within the first hundred milliseconds after the onset of the human voice (vs. matched synthesized sounds) result- ing in the activation of neural patterns thought to be involved in constructing the acoustic object, and (ii) later activity (after 1000 ms), as part of the mechanism driven by the emotional response related to the stimuli. Finally, on the basis of Eitan et al.’s study (2013), we predicted the observation of hemispheric specialization for emotional prosody processing in favor of the right STN.
2. Material and methods
2.1. Patients and surgery (Table 1)
Fifteen patients took part in the study. All of them met the clin- ical criteria of the Parkinson’s UK Brain Bank for idiopathic PD (Hughes, Daniel, Kilford, & Lees, 1992). The patients, all with med- ically intractable PD, underwent bilateral STN DBS at Rennes University Hospital (France); two patients underwent right unilat- eral STN DBS. Standard selection and exclusion criteria for surgery were applied to all of them (Welter et al., 2002). In particular, brain atrophy was excluded on the basis of a preoperative MRI.
Three months prior to the STN DBS surgery and the electrophys- iological recordings, all patients were assessed in accordance with the Core Assessment Program for Intracerebral Transplantation (Langston et al., 1992) and scored on the Unified Parkinson’s Dis- ease Rating Scale I–IV (Fahn & Elton, 1987), Hoehn and Yahr scale (Hoehn & Yahr, 1967), and Schwab and England scale (Schwab &
England, 1969). Patients were assessed on and off dopa both before and after surgery. In addition, all the patients were neuropsycho- logically assessed to rule out cognitive impairments (Péron et al., 2010, 2011). The neuropsychological battery included the Mattis Dementia Rating Scale (Mattis, 1988) and a series of tests assessing frontal executive functions, including Nelson’s modified version of the Wisconsin Card Sorting Test (Nelson, 1976), the Trail Making Test (Reitan, 1958), the Categorical and Literal Fluency Test (Cardebat, Doyon, Puel, Goulet, & Joanette, 1990), and the Stroop Test (Stroop, 1935). None of the patients included in this study had dementia (seeTable 1). As it has been previously shown that depression can influence emotional prosody recognition (Péron et al., 2011), we controlled for this variable and observed that the patients did not score above the cut-off on the depression scale (seeTable 1). Finally, none of the patients included in the study wore hearing aids or had a history of tinnitus, and all patients were assessed by means of a standard audiometric screening procedure (AT-II-B audiometric test); two patients were excluded from the study because of presbycusis with bilateral hearing loss that was more marked at higher frequencies. Following surgery, patients were followed up clinically by a movement disorder specialist.
The study was approved by the ethics committee of Rennes University Hospital (approval number IDRCB: 2011-A00392-39).
After providing participants with a complete description of the study, written informed consent was obtained from each, and the
study was conducted in accordance with the Declaration of Helsinki.
The macroelectrodes used were the Medtronic model 3389 (Medtronic Neurological Division, Minneapolis, MN, USA) with four platinum-iridium cylindrical surfaces (1.27 mm in diameter and 1.5 mm in length) and a contact-to-contact separation of 0.5 mm. Contact 0 was the most ventral and caudal, and Contact 3 the most dorsal and rostral. The surgical procedure consisted first, of the attachment to the patient’s head of a stereotactic Lek- sell frame under local anesthesia, and then the implantation of bilateral quadripolar DBS electrodes in the postero-lateral part of the STN in a single surgical session. The overall methodology was similar to that previously described by Benabid et al. (2000). A three-dimensional computed tomography (CT) scan of the brain with the Leksell frame was performed at the onset of the surgical procedure and was registered to the preoperative MRI in order to calculate the stereotactic coordinates for positioning the two selected electrode contacts (one on the left and one on the right).
The intended coordinates at the tip of Contact 0 were 10–12 mm from the midline, 0–3 mm behind the midcommissural point, and 3–5 mm below the anterior commissure (AC)-posterior com- missure (PC) line. During the operation, the final course and depth of the electrode were determined by the best effect obtained on rigidity with no side effect and at the lowest voltage. A three- dimensional CT brain scan performed a few days later confirmed the position of the electrodes. The pre- and postoperative image segmentation and registration workflow, the anatomical ParkMe- dAtlis template used as an anatomical reference (Haegelen et al., 2013), and the electrode segmentation were performed by using pyDBS software (D’Albis et al., 2015).
2.2. Experimental setup 2.2.1. Stimuli (Fig. S1)
The vocal (prosodic) stimuli consisted of two pseudosentences spoken with different emotional prosodies (‘‘ne kali bam sud molen!” and ‘‘kun se mina lod belam?”; mean duration = 1642 ms, range = 854–2788 ms) extracted from a previously validated data- Table 1
Average (±SD) of the clinical, motor, neuropsychological and psychiatric scores before STN DBS (preoperative condition; baseline) displayed by the thirteen patients with Parkinson’s disease included in the ERPs analyses.
N = 13 PD patients
Age (yrs) 53.8 (±7.7)
Education level (yrs) 11.5 (±3.1)
Sex N = 5 F; N = 8 M
Disease lateralization N = 4 R; N = 9 L
Disease duration (yrs) 8.0 (±2.9)
UPDRS-III ON dopa 15.7 (±10.2)
UPDRS-III OFF dopa 37.0 (±10.3)
S&E (%) ON dopa 86.1 (±15.0)
S&E (%) OFF dopa 69.2 (±22.5)
H&Y ON dopa 0.9 (±1.1)
H&Y OFF dopa 2.3 (±0.9)
LED (at the time of LFP recordings) 1047.1 (±692.0)
MDRS (/144) 140.0 (±3.2)
Stroop test - Interference 0.12 (±7.4)
TMT B-A (s) 50.4 (±40.3)
Categorical verbal fluency (20) 30.9 (±11.4)
Phonemic verbal fluency (20) 17.8 (±6.5)
MCST - Number of categories (/6) 5.6 (±0.9)
MCST - Number of errors 4.9 (±5.8)
MCST - Number of perseverative errors 1.5 (±1.9)
MADRS 2.5 (±1.9)
Abbreviations: F: female; M: male; R: right; L: left; UPDRS: United Parkinson’s Disease Rating Scale; S&E: Schwab & England; H&Y: Hoehn & Yahr; LED: levodopa- equivalent dose; MDRS: Mattis Dementia Rating Scale; TMT: Trail Making Test;
MCST: Modified Wisconsin Card Sorting Test.
base, the GEneva Multimodal Emotion Portrayals corpus (Banziger
& Scherer, 2010). Alongside these prosodic stimuli (anger, happi- ness and neutral), we played synthesized stimuli, built from the original emotional and neutral sounds, in order to control for the temporal dynamics of energy andf0. These two basic acoustic fea- tures are known to be the most correlated with emotional prosody judgments (e.g.,Banse & Scherer, 1996; Grandjean et al., 2006). The first type of synthetic stimulus (synthesizedintensity) consisted of a section of white noise, to which the intensity contour of the orig- inal stimulus was applied. The second synthetic stimuli (denomi- nated as ‘‘synthesized f0”) was a sine wave with constant amplitude. Its frequency is variable and corresponds to the detectedf0 of the original recording. Both synthetic stimuli had the same duration as in the original recordings. Examples of stim- uli in each of the prosodic conditions (anger, happiness, neutral) are provided inSupplemental Fig. S1 (oscillograms and spectro- grams). All the sounds were matched for mean energy to avoid loudness effects. Three blocks were constructed, featuring the dif- ferent kinds of stimuli in pseudorandom order (no more than three times for the same experimental condition). Each block contained 20 trials featuring anger stimuli, 20 trials featuring happiness stim- uli, and 20 trials featuring neutral stimuli, as well as 15 synthe- sized intensity stimuli, 15 synthesizedf0stimuli, and 1 section of white noise at the beginning (first stimulus) with a gradual attack to accustom the patients to the sounds. Each block contained a dif- ferent list of stimuli. In each prosodic condition, we controlled for the pseudo-sentence being pronounced and the sex of the actor who pronounced the utterances: half the stimuli were pronounced by a female actor, and half the stimuli consisted of the pseudo- sentence ‘‘ne kali bam sud molen!” The total duration of each block was10 min, and there was a break between each one (i.e., a total of two breaks during the experiment). Each block contained pairs of identical stimuli, representing 10% of the stimuli, which were again ordered pseudorandomly. Their usefulness is explained in Section2.2.3(Task).
2.2.2. Trial (Fig. 1)
An example of a trial is provided inFig. 1. In each trial, in order to avoid expectancy effects, we varied the duration of the interval between the onset of the fixation cross and the onset of the audi- tory stimulus. In other words, the presentation of each auditory stimulus was preceded by a silent portion of pseudorandom dura- tion, ranging from 50 to 250 ms, the so-called jitter (see Jitter 1;
Fig. 1). After the offset of the sound, we also included a silent por- tion ranging from 3000 to 3500 ms (see Jitter 2 + Silence;Fig. 1). In order to avoid the offset of the sound and the offset of the fixation cross being synchronous, we varied the duration of the interval between these two offsets (see Jitter 2;Fig. 1). Finally, in order to minimize any retinal afterimage, we ensured that the color of the fixation cross did not contrast too greatly with the color of the desktop background.
2.2.3. Task (Fig. 1)
For each trial, the patients were asked to keep their eyes open and relaxed. They were told that they would hear meaningless speech uttered by male and female actors, as well as synthesized sounds. The loudness intensity was adjusted for each patient according to their hearing threshold at the beginning of the exper- iment. Participants were asked to focus on these auditory stimuli and to press a button whenever they heard two identical stimuli in a row (see one-back trial;Fig. 1). These one-back trials repre- sented only 10% of the totality of the trials and were excluded from the analyses. All the stimuli were played bilaterally through in-ear stereo headphones (Sennheiser, CX 2.00Ò). This one-back task was administered to ensure that the patients were attentive to the sounds. Absolute silence was required for the test. Prior to the task,
a pad with a push button was placed beneath the patients’ fingers.
This pad generated a signal that was simultaneously recorded by the equipment. We registered the presentation of the various pro- sodic and synthesized stimuli onto the resulting electrophysiolog- ical recordings. The hand (right or left) used to press the button was randomized across participants. Half of the patients used their dominant hand and the other half their nondominant hand. How- ever, trials in which the patient gave a motor response were also excluded.
2.3. Recordings
All the patients were studied 2 days postoperatively in the interval between DBS electrode implantation and subsequent con- nection to a subcutaneous stimulator. They were all taking their antiparkinsonian medication when they were assessed. Subthala- mic LFPs were recorded through a g.BSampÒ(g.tec Medical Engi- neering, Schiedlberg, Austria) biosignal amplifier connected to a PowerLabÒ16/35 (ADInstruments, Dunedin, New Zealand) system.
Deep brain activity was recorded bipolarly from two adjacent con- tacts of each DBS electrode (three bipolar derivations 0–1, 1–2, and 2–3, Contact 0 being the most distal contact), amplified, and sam- pled at a common rate of 1000 Hz. Signals were monitored online by using LabchartÒ(ADInstruments) software.
2.4. Data analysis
Data analysis was performed by using C++ software developed in-house by Lucas Tamarit, an engineer at the Swiss Center for Affective Sciences in Geneva (Switzerland); Cartool software devel- oped by Denis Brunet (brainmapping.unige.ch/cartool); and the Matlab FieldTrip toolbox (http://fieldtrip.fcdonders.nl/;
Oostenveld, Fries, Maris, & Schoffelen, 2011).
LFP signals were filtered offline with a 0.1–30 Hz bandpass FIR filter. Stimulus-locked epochs were then selected as follows: as the audio stimuli had a minimum duration of 854 ms, and the com- bined silent portions between them had a minimum duration of 3100 ms, we chose to create epochs starting 200 ms before the onset of the stimulus and ending 3700 ms post-onset. A baseline correction was applied to each individual epoch by subtracting the average level measured during the 200 ms preceding stimulus onset. These epochs served as the input for all subsequent analyses.
The selection of the STN contact pairs for the analyses was based on their anatomical position: at least one of the two contacts in the contact pair was located within the STN (N= 24 contact pairs from both sides in 13 patients, 2 right unilateral, as revealed ear- lier; 2 patients were excluded from the LFP analyses because of basic auditory impairment). For the data analyses, we selected the first level of recordings for each patient: dipole 0–1 in both the left and the right STN. Epochs containing muscular artifacts or excessive noise were removed from the data by means of visual inspection (5% rejection). One-back trials and trials in which the patient gave a motor response were also excluded. ERPs were derived by averaging stimulus-locked epochs in 3074 trials: 2168 trials for the human voice (613 for anger, 791 for happiness, 764 for neutral) and 906 trials for the synthesized stimuli. Finally, to remove some of the noise in the data, and to avoid excessive wig- gliness in the analysis, we first smoothed the epochs with a sliding window of 30 ms.
We performed several planned comparisons of the ERPs of the different conditions (see above). As the data were not binned (i.e., not average in time), analyses were carried out on all 3900 time points (i.e., one every 1 ms) of the epochs. For each of the 3900 time points, we computed a Student t statistic comparing two conditions (e.g., anger vs. neutral), but did not compute the
correspondingpvalue. We applied a very stringent approach to multiple testing corrections in order to control for the familywise error rate (FWER) over the 3900 tests to a level of 5%. To this end, we used the so-calledcluster method(Maris & Oostenveld, 2007), which has the additional advantage of being a nonparametric method and therefore does not rely on Gaussian assumptions.
This method clusters adjacent times for which the Student t statistic is above a given threshold and combines them to obtain a statistic for the cluster. To test all the time points simultane- ously, we constructed N (typically = 1000) surrogate conditions by shuffling (permuting) the epochs of the two conditions to compare the values of these cluster statistics and obtain apvalue for each cluster. Only clusters withpvalues smaller than 5% are reported in the Results section. It has been shown that these regions can be deemed significant if they are below a FWER threshold of 5% (Maris & Oostenveld, 2007). To take into account that the epochs are measured on different patients, we standard- ized the epochs by participant and we modified the Fieldtrip function so that epochs were permuted only within participants, in accordance with the principle that only similar epochs can be permuted. We used the 0.8 quantile as a threshold and the
wcm cluster statistic (with the default power value of two).
One of the great advantages of the permutation method is the freedom to choose the statistic. To check the sensitivity of our results, we therefore repeated the analysis, first without the stan- dardization and the per participant restriction for the permuta- tion and second by replacing the Student t statistic with a robust version based on 10% trimmed means. The results (not shown) were close to the results based on the Studentt, showing that they were not influenced by a few outlying epochs, but were actually quite robust.
Planned comparisons across the patients were performed by means of the nonparametric pairwise permutation tests described above, with the corrected clusterpvalue, considering the left and the right STN separately. We conducted two kinds of comparisons.
First, we explored differences in the patterns of electrophysiologi- cal activity between the neutral, angry, and happy prosodies (human voice) and the synthesized (nonvoiced) stimuli. In order to investigate the emotional effects, we then looked for differences in electrophysiological activity in response to the angry and happy prosodies versus the neutral prosody, as well as in response to the angry versus happy prosodies.
Fig. 1.Trial examples and task. Legend: Jitter 1 = pre-onset silent portion of pseudorandom duration lasting between 50 and 250 ms in order to avoid an expectancy effect;
Sound = auditory stimulus lasting between 854 and 2788 ms; Jitter 2 = post-offset silent portion of pseudorandom duration lasting between 50 and 500 ms in order to avoid synchronicity between the offset of the sound and the offset of the fixation cross; Jitter 2 + Silence = post-offset silent portion of pseudorandom duration lasting between 3000 and 3500 ms. For each trial, the patients were told that they would hear meaningless speech uttered by male and female actors, as well as synthesized sounds. They were asked to focus on these auditory stimuli and to press a button whenever they heard two identical stimuli in a row (as in the one-back trial example presented here). All the stimuli were played bilaterally through in-ear stereo headphones. This one-back task was administered to ensure that the patients were attentive to the sounds. The hand (right or left) used to press the button was randomized across participants.
3. Results
3.1. Anatomical location (Fig. 2)
As we selected the first level of recordings for each patient (dipole 0–1 in both the left and the right STN), we calculated the mean contact for each patient between Contact 0 and 1, and then calculated the average coordinates of this mean contact between all patients in a referential space, the ParkMedAtlis template (Haegelen et al., 2013). The mean (Talairach) coordinates of this mean contact with the AC as the origin of the coordinates were as follows: left STN, y =10.9 ± 1.3 mm in the antero-posterior direction, x =15.2 ± 1.6 mm in the lateral direction, z =5.3 ± 2.0 mm under a line passing through the AC and the PC line; right STN, y = 10.6 ± 1.4 mm, x =14.8 ± 1.6 mm, z =3.77 ± 2.2 mm. Fig. 2provides a view of the location of the mean contacts for the 13 patients, showing that the left and the right mean contacts were inside the STN. Individual coordinates are also provided inSupplemental Table S2.
3.2. Behavioral results
In the one-back task, all patients had a correct hit rate of 100%
and an average of 4.2% false positive responses (resulting in exclu- sion of ERP analyses), with no missed responses.
3.3. Event-related potentials
3.3.1. Human voice versus synthesized stimuli (Fig. 3)
As emotional prosody processing is subtended by human voice processing, and as previously explained in Section2.4(Data Anal- ysis), we first ran a planned contrast, comparing the human voice (anger, happiness, and neutral) with the synthesized stimuli (i.e., synthesized intensity and synthesized f0) (see Subsection 2.2.1 (Stimuli) of the Material and Methods section andSupplemental Material S1).
As shown inFig. 3, results revealed significantly greater activity in the human voice condition than in the synthesized stimuli con- dition, with early activity in theleft and the rightSTN (left STN:
300 to 400 ms and 450 to 700 ms post-onset, corrected p< 0.05; right STN: 500 to 600 ms post-onset, corrected p< 0.05).
3.3.2. Emotional effects
3.3.2.1. Angry versus neutral (Fig. 4). As shown inFig. 4, results indi- cated significantly greater activity in the anger condition than in the neutral one (early activity200 to300 ms and later activity 850to950 ms post-onset, correctedp< 0.05) in therightSTN.
No significant difference was observed between the two conditions in theleftSTN.
3.3.2.2. Happy versus neutral (Fig. 5). As shown inFig. 5, when we analyzed happy minus neutral prosodies, results revealed a signif- icant difference between the two conditions in therightSTN (200 to 300 ms and 3250 to 3350 ms post-onset, corrected p< 0.05). No significant difference was observed between the two conditions in theleftSTN.
3.3.3. Lateralization effects (right versus left STN comparisons) When we compared electrophysiological activity in the right versus theleftSTN, in each of the anger and happiness conditions, we did not observe significant differences.
4. Discussion
The aim of the present study was to explore the influence of positive and negative vocal emotions (i.e., emotional prosody) on the electrophysiological activity of the STN in humans. We com- pared the electrophysiological activity (LFPs) of the left and the right STN in 13 PD patients in response to angry, happy, and neu- tral prosodies, as well as to nonhuman synthesized stimuli. We were thus able to perform two major types of analyses. First, we explored the different patterns of electrophysiological activity in response to the human voice (neutral, anger, and happy prosodies) versus the synthesized stimuli (intensity contour applied to white noise, andf0contour applied to a pure sine oscillator; seeSupple- mental Material S1). Second, in order to investigate the emotional effects, we looked for differences in electrophysiological activity in response to the angry and happy prosodies versus the neutral prosody.
Before we set out our results and draw any inferences from them, it is important to acknowledge that, as in many clinical stud- ies of this nature, the sample size was small (N= 13 patients included in the LFP analyses, 24 STN) which represents a limita- tion. In order to address this point, and although this is not rou- tinely done in the literature, we applied a very stringent statistical approach to multiple testing correction to control for FWER. In addition, we investigated the sensitivity of our results to one specific participant or to a few epochs because it is known that thettest is not robust against outliers. We thus ran a robust analysis (with 10% trimmed means). As this analysis yielded simi- lar results to those based on thetstatistic, we are confident that our findings are not due just to one participant or to outlying epochs.
First of all, we would like to stress a clear event-related compo- nent in all our average ERPs, demonstrating a sensitivity of the STN for sound processing in general. This ERP complex is characterized by positivity at80 ms after the onset of the sound, followed by negativity at150 ms, itself followed by a positive fluctuation at 250 ms. The first significant difference in our experimental Fig. 2.Anterior view showing the basal ganglia segmentation and the mean contacts inside the subthalamic nucleus on the right and the left sides for 13 patients with Parkinson’s disease. These mean left and right contacts were inside the STN. L = left; R = right; amygdala (light purple); putamen (dark purple);
thalamus (yellow); subthalamic nucleus (orange); substantia nigra (grey); red nucleus (red); lateral pallidum (green); medial pallidum (dark orange); ventricles (dark blue). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 3.Stimulus-lockedleft(top panel) andright(bottom panel) STN ERPs in the human voice (black) (n= 2168 trials) and synthesized stimuli (red) (n= 906 trials) response conditions from the 13 patients and from the first level of recording for each patient (dipole 0–1). The vertical black line at 0 ms indicates the onset of the auditory stimulus.
Time regions where the difference between the two conditions is statistically significant, controlling for FWER at 5%, are shaded in grey. Filtering = 0.1-Hz high-pass and 30- Hz low-pass FIR filters, baseline correction ranging from 0 to 200 ms, temporal smoothing with a sliding window of 30 ms. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 4.Stimulus-lockedleft(top panel) andright(bottom panel) STN ERPs in the anger (black) (n= 613 trials) and neutral (red) (n= 764 trials) response conditions from the 13 patients and from the first level of recording for each patient (dipole 0–1). The vertical grey line at 0 ms indicates the onset of the auditory stimulus. Time regions where the difference between the two conditions is statistically significant, controlling for FWER at 5%, are shaded in grey. Filtering = 0.1-Hz high-pass and 30-Hz low-pass FIR filters, baseline correction ranging from 0 to 200 ms, temporal smoothing with a sliding window of 30 ms. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
design (see below) occurs within this third component. Further research is needed to better understand the functional role of these early neural responses that were not modulated by our experimen- tal manipulation.
As far as the human voice versus non-voiced stimuli compar- isons are concerned, we observed a specific modulation of the STN in response to the human voice. There was significantly greater activity in the human voice condition than in the synthe- sized stimuli condition in both therightand theleftSTN, occurring early in theleft(300 to400 ms post-onset, correctedp< 0.05) and then almost simultaneously in both therightand theleft(right STN: 500 to 600 ms post-onset, corrected p< 0.05; left STN:
450 to700 ms post-onset, correctedp< 0.05) (Fig. 3). Our com- parison of human voices versus synthesized stimuli enabled us to measure the STN’s specific activity in response to human voices minus the variations in f0 and energy. In other words, when we performed this comparison, we could test the STN’s specific activ- ity in response to the remainingspectralcomponents of the human voice signal (and other related processes) by controlling forf0and intensity dynamics using synthesized sounds. The involvement of cortical (auditory) structures in the processing of such spectral modulations of speech has been reported in several studies (e.g., Leaver & Rauschecker, 2010; Overath, Kumar, von Kriegstein, &
Griffiths, 2008). However, to the best of our knowledge, apart from the involvement of these structures in the temporal organization of events (Kotz & Schwartze, 2010), the functional specialization of the BG and, more specifically, the STN has never before been reported in this process. It would be worthwhile for future studies to investigate the functional specialization and integration of these subcortical structures regarding the spectral components of voice processing in greater detail. Nonetheless, and based on Graybiel’s model (for an exhaustive description of her model, seeGraybiel, 2008), this result is not surprising. This model proposes that the BG may be critically involved in constructingperformance units of sequence representations, also calledchunks. The proposition is that items of information can be organized more efficiently by recoding
them to form packages (chunks) that, once learned, can be pro- cessed more automatically. The nature of these items can be emo- tional, as we suggested recently (Péron et al., 2013), as well as motor and cognitive (for e.g., vocal).
As far as the investigation of the emotional effects is concerned, we observed a specific modulation of the STN in response to both angry and happy prosodies only in therightSTN. We observed two different temporal dynamics: (i) right STN ERPs occurring shortly after onset and at the same time for angry and for happy prosodies (Figs. 4 and 5, bottom panels;200 to300 ms post-onset, cor- rectedp< 0.05); and (ii) later activity for angry prosody at850 to 950 ms post-onset (corrected p< 0.05) and at 3250 to 3350 ms post-onset (correctedp< 0.05) for happy prosody. No significant difference was observed between the two conditions in theleftSTN (Figs. 4 and 5, top panels).
In accordance with our hypotheses, the present results seem to confirm the STN’s functional involvement in emotional prosody, whether the valence is positive (happy voices) or negative (angry voices). In addition, and as we had predicted, the STN ERP modula- tions observed in response to vocal emotions occurred both early and later during emotion processing. As indicated in the Introduc- tion, studies in the visual modality have reported event-related desynchronization between 1000 and 2000 ms post-onset (Brucke et al., 2007; Huebl et al., 2011; Kühn et al., 2005). In Kühn and colleagues’ study, as well as in Huebl and colleagues’
study, these results were obtained for both positive and negative IAPS stimuli, whereas in Brücke and colleagues’ study, only posi- tive images were presented. To the best of our knowledge, the early effects we obtained in response to anger and happy prosodies are new empirical evidence for the involvement of the STN at an early stage of emotion processing. One possible explanation for these potentially modality-driven differences is the existence of specific timing patterns for the visual versus auditory modalities.
Taken together, the present results supportPéron et al. (2013) posit that the STN plays a meta-role and exhibits supramodal spe- cialization in emotional processes. According to this model, the Fig. 5.Stimulus-lockedleft(top panel) andright(bottom panel) STN ERPs in the happiness (black) (n= 791 trials) and neutral (red) (n= 764 trials) response conditions from the 13 patients and from the first level of recording for each patient (dipole 0–1 for both electrodes). The vertical grey line at 0.2 s indicates the onset of the auditory stimulus.
Time regions where the difference between the two conditions is statistically significant, controlling for FWER at 5%, are shaded in grey. Filtering = 0.1-Hz high-pass and 30- Hz low-pass FIR filters, baseline correction ranging from 0 to 200 ms, temporal smoothing with a sliding window of 30 ms. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
STN may thus be involved in the multiple stages of emotion pro- cessing. In addition, the finding in the present results of the STN’s functional role in processing vocal emotions reinforces the hypoth- esis whereby the BG and, in this case, the STN, are sensitive to both the temporal and structural organization of events.Péron et al.
(2013)model ascribes to the STN the intrinsic functional role of neural rhythm organizer at the cortical and subcortical levels in emotion processing, which predicts its sensitivity to rhythm. In the context of vocal emotion recognition, for example, the STN would act as a marker for the transiently connected neural net- work (i.e., amygdala, auditory cortices, inferior frontal gyri, and orbitofrontal cortices) that subserves emotional prosody process- ing. If the co-activation across different neuronal populations is recurrent or functionally important, the BG-mediated synchroniza- tion presumably increases the weight of the synaptic connections within the cortical or subcortical neural network.
In addition to these responses, we also observed effects pointing to the STN’s hemispheric specialization in emotional prosody pro- cessing in accordance with Eitan et al. (2013). When we tested simple effects by analyzing therightand theleftSTN separately, we were not able to reject the null hypothesis for the comparisons in theleftSTN (Figs. 4 and 5, top panels), whereas we were able to do so in therightSTN for the angry versus neutral and happiness versus neutral prosodies (Figs. 4 and 5, bottom panels). In order to investigate these results in greater detail, we performed analy- ses comparing the electrophysiological activity of therightversus leftSTN for anger and for happiness. The analysis did not reveal any significant difference in electrophysiological activity between therightand theleftSTN for either of the conditions we tested.
However, given that we were not able to test these interaction effects directly, the present results can still be regarded as reflect- ing an asymmetry in terms of right/left lateralization interacting with our experimental conditions taken together. Further studies are needed to carry out a systematic investigation of this obviously complex pattern, especially bearing in mind that symptoms were predominantly left-sided for 9 of the 13 patients included in the analyses, thus limiting the inferences that can be drawn from these results.
That being said, it is worth noting that models of emotional pro- sody perception allow for right hemispheric specialization at many stages in the process (see,Witteman et al., 2012 for a review of these models). Two recent meta-analyses, one of the lesion litera- ture (Witteman et al., 2011), the other of the neuroimaging litera- ture (Witteman et al., 2012), also seem to confirm this hypothesis, pointing to the conclusion that emotional prosody perception takes place mainly in therightfrontotemporal network, but also involves therightBG. The meta-analysis of the lesion literature found statis- tically robust evidence of a greater deterioration in emotional pro- sodic performance following right rather than left hemispheric BG damage (Witteman et al., 2011). The meta-analysis of neuroimag- ing literature reported clusters in the right medial globus pallidus and caudate body when emotional prosody stimuli were compared with neutral stimuli or synthesized speech (Witteman et al., 2012).
Taken together, the present results appear to indicate possible right hemispheric specialization for emotional prosody processing in the BG. Focusing on the STN, apart from the study byEitan et al.
(2013), the only information concerning this basal ganglion’s potential hemispheric specialization in emotional (prosody) pro- cessing comes from an fMRI study in healthy participants under- taken byPéron, Frühholz, et al. (2016). This study yielded results that appear to contradict the right dominance hypothesis, as it was theleftSTN that exhibited activity during the processing of emotional (angry voices) versus neutral stimuli. This activity was observed only in a condition in which the listeners focused on a nonemotional feature of the voice (i.e., the speaker’s sex; sex iden- tification task), instead of concentrating on emotional vocal fea-
tures (i.e., emotion discrimination; prosody task). In this context, we can hypothesize that hemispheric specialization differs accord- ing to whether it is driven by the stimulus and/or by the task, espe- cially when the task requires the inhibition of prepotent responses (as was the case in Péron, Frühholz, and colleagues’ study). This hypothesis of differential (left vs. right) functional roles played by the STN in emotional (but also cognitive and motor) processing deserves to be explored more thoroughly in future studies, given its fundamental, but also clinical, relevance.
In conclusion, we found both early and late modulation of STN activity in response to the human voice, as well as specific early and later activity in response to negative and positive emotions conveyed by voices. These results confirm the hypothesis that the STN is involved in emotion processing, irrespective of stimulus valence (positive or negative) and sensory modality (e.g., visual or auditory). Future studies featuring a larger number of participants, together with neuroimaging studies in healthy participants, are needed to confirm our findings.
Disclosure
The authors report no conflicts of interest. The study was car- ried out at the Neurology Unit of Pontchaillou Hospital (Rennes University Hospital, rue Henri Le Guilloux, 35033 Rennes, France;
Prof. Marc Vérin). Data acquisition (externalization and acquisition per se) was funded by PHRC-IR Grant No. IDRCB: 2011-A00392-39 (Prof. Marc Vérin). The first author (Dr. Julie Péron) was funded by Swiss National Foundation Grant No. 105314_140622 (Prof. Didier Grandjean and Dr. Julie Péron), and NCCR Affective Sciences was funded by Swiss National Foundation Project No. 202 – UN7126 (Prof. Didier Grandjean). The funders had no role in data collection, discussion of content, preparation of the manuscript, or decision to publish.
Acknowledgements
The study was carried out at the Neurology Unit of Pontchaillou Hospital (Rennes University Hospital, rue Henri Le Guilloux, 35033 Rennes, France; Prof. Marc Vérin). Data acquisition (externalization and acquisition per se) was funded by PHRC-IR Grant No. IDRCB:
2011-A00392-39 (Prof. Marc Vérin). The first author (Dr. Julie Péron) was funded by Swiss National Foundation Grant No.
105314_140622 (Prof. Didier Grandjean and Dr. Julie Péron), and NCCR Affective Sciences was funded by Swiss National Foundation Project No. 202 – UN7126 (Prof. Didier Grandjean). The funders had no role in data collection, discussion of content, preparation of the manuscript, or decision to publish. The Cartool software (b rainmapping.unige.ch/cartool) was programmed by Denis Brunet, from the Functional Brain Mapping Laboratory, Geneva, Switzer- land, and supported by the Center for Biomedical Imaging (CIBM) of Geneva and Lausanne. We would like to thank the patients for contributing their time to this study. We are also grateful to Chris- tophe Mermoud for his help in setting up the acquisition system and to Elizabeth Wiles-Portier for revising the English style.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.bandl.2016.12.
003.
References
Bach, D. R., Grandjean, D., Sander, D., Herdener, M., Strik, W. K., & Seifritz, E. (2008).
The effect of appraisal level on processing of emotional prosody in meaningless speech.Neuroimage, 42(2), 919–927.
Banse, R., & Scherer, K. R. (1996). Acoustic profiles in vocal emotion expression.
Journal of Personality and Social Psychology, 70(3), 614–636.
Banziger, T., & Scherer, K. R. (2010). Introducing the Geneva Multimodal Emotion Portrayal (GEMEP) corpus. In T. Banziger, K. Scherer, & E. Roesch (Eds.),A blueprint for an affectively competent agent: Cross-fertilization between emotion psychology, affective neuroscience, and affective computing. Oxford: Oxford University Press.
Belin, P., Fillion-Bilodeau, S., & Gosselin, F. (2008). The Montreal affective voices: A validated set of nonverbal affect bursts for research on auditory affective processing [Research Support, Non-U.S. Gov’t Validation Studies]. Behavior Research Methods, 40(2), 531–539.
Belin, P., & Zatorre, R. J. (2000). ’What’, ’where’ and ’how’ in auditory cortex.Nature Neuroscience, 3(10), 965–966.
Benabid, A. L., Koudsie, A., Benazzouz, A., Fraix, V., Ashraf, A., Le Bas, J. F., ... Pollak, P.
(2000). Subthalamic stimulation for Parkinson’s disease.Archives of Medical Research, 31(3), 282–289.
Benedetti, F., Colloca, L., Lanotte, M., Bergamasco, B., Torre, E., & Lopiano, L. (2004).
Autonomic and emotional responses to open and hidden stimulations of the human subthalamic region.Brain Research Bulletin, 63(3), 203–211.
Boersma, P., & Weenink, D. (2013). Praat: Doing phonetics by computer Version 5.3.51 Retrieved 2 June 2013 from<http://www.praat.org/>.
Bruck, C., Kreifelts, B., & Wildgruber, D. (2011). Emotional voices in context: A neurobiological model of multimodal affective information processing [Review]. Physics of Life Reviews, 8(4), 383–403. http://dx.doi.org/10.1016/
j.plrev.2011.10.002.
Bruck, C., Wildgruber, D., Kreifelts, B., Kruger, R., & Wachter, T. (2011). Effects of subthalamic nucleus stimulation on emotional prosody comprehension in Parkinson’s disease.PLoS ONE, 6(4), e19140.http://dx.doi.org/10.1371/journal.
pone.0019140.
Brucke, C., Kupsch, A., Schneider, G. H., Hariz, M. I., Nuttin, B., Kopp, U., ... Kuhn, A. A.
(2007). The subthalamic region is activated during valence-related emotional processing in patients with Parkinson’s disease. European Journal of Neuroscience, 26(3), 767–774.
Buzsaki, G., Anastassiou, C. A., & Koch, C. (2012). The origin of extracellular fields and currents–EEG, ECoG, LFP and spikes [Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov’t Research Support, U.S. Gov’t, Non-P.H.S.
Review]. Nature Reviews Neuroscience, 13(6), 407–420. http://dx.doi.org/
10.1038/nrn3241.
Cardebat, D., Doyon, B., Puel, M., Goulet, P., & Joanette, Y. (1990). Formal and semantic lexical evocation in normal subjects. Performance and dynamics of production as a function of sex, age and educational level.Acta Neurologica Belgica, 90(4), 207–217.
Courtin, J., Karalis, N., Gonzalez-Campo, C., Wurtz, H., & Herry, C. (2014). Persistence of amygdala gamma oscillations during extinction learning predicts spontaneous fear recovery [Research Support, Non-U.S. Gov’t].Neurobiology of Learning and Memory, 113, 82–89.http://dx.doi.org/10.1016/j.nlm.2013.09.015.
D’Albis, T., Haegelen, C., Essert, C., Fernandez-Vidal, S., Lalys, F., & Jannin, P. (2015).
PyDBS: An automated image processing workflow for deep brain stimulation surgery.International Journal of Computer Assisted Radiology and Surgery, 10(2), 117–128.http://dx.doi.org/10.1007/s11548-014-1007-y.
Delplanque, S., N’Diaye, K., Scherer, K. R., & Grandjean, D. (2007). Spatial frequencies or emotional effects? A systematic measure of spatial frequencies for IAPS pictures by a discrete wavelet analysis.Journal of Neuroscience Methods, 165(1), 144–150.
Eitan, R., Shamir, R. R., Linetsky, E., Rosenbluh, O., Moshel, S., Ben-Hur, T., ... Israel, Z.
(2013). Asymmetric right/left encoding of emotions in the human subthalamic nucleus. Frontiers in Systems Neuroscience, 7, 69.http://dx.doi.org/10.3389/
fnsys.2013.00069.
Escera, C., Leung, S., & Grimm, S. (2014). Deviance detection based on regularity encoding along the auditory hierarchy: Electrophysiological evidence in humans [Research Support, Non-U.S. Gov’t Review].Brain Topography, 27(4), 527–538.http://dx.doi.org/10.1007/s10548-013-0328-4.
Ethofer, T., Anders, S., Erb, M., Herbert, C., Wiethoff, S., Kissler, J., ... Wildgruber, D.
(2006). Cerebral pathways in processing of affective prosody: A dynamic causal modeling study.Neuroimage, 30(2), 580–587.
Fahn, S., & Elton, R. (1987). UPDRS development committee. Unified Parkinson’s disease rating scale.Recent Development in Parkinson’s Disease, 2, 153–163.
Fawcett, A. P., Dostrovsky, J. O., Lozano, A. M., & Hutchison, W. D. (2005). Eye movement-related responses of neurons in human subthalamic nucleus.
Experimental Brain Research, 162(3), 357–365.
Frühholz, S., Ceravolo, L., & Grandjean, D. (2012). Specific brain networks during explicit and implicit decoding of emotional prosody.Cerebral Cortex, 22(5), 1107–1117.http://dx.doi.org/10.1093/cercor/bhr184.
Frühholz, S., & Grandjean, D. (2013). Processing of emotional vocalizations in bilateral inferior frontal cortex [Research Support, Non-U.S. Gov’t Review].
Neuroscience and Biobehavioral Reviews, 37(10 Pt 2), 2847–2855.http://dx.doi.
org/10.1016/j.neubiorev.2013.10.007.
Grandjean, D., Banziger, T., & Scherer, K. R. (2006). Intonation as an interface between language and affect.Progress in Brain Research, 156, 235–247.
Grandjean, D., Sander, D., Lucas, N., Scherer, K. R., & Vuilleumier, P. (2008). Effects of emotional prosody on auditory extinction for voices in patients with spatial neglect. Neuropsychologia, 46(2), 487–496. http://dx.doi.org/10.1016/j.
neuropsychologia.2007.08.025. S0028-3932(07)00307-7 [pii].
Grandjean, D., Sander, D., Pourtois, G., Schwartz, S., Seghier, M. L., Scherer, K. R., &
Vuilleumier, P. (2005). The voices of wrath: Brain responses to angry prosody in meaningless speech.Nature Neuroscience, 8(2), 145–146.
Graybiel, A. M. (2008). Habits, rituals, and the evaluative brain [Research Support, N.
I.H., Extramural Research Support, Non-U.S. Gov’t Research Support, U.S. Gov’t, Non-P.H.S. Review].Annual Review of Neuroscience, 31, 359–387.http://dx.doi.
org/10.1146/annurev.neuro.29.051605.112851.
Haegelen, C., Coupe, P., Fonov, V., Guizard, N., Jannin, P., Morandi, X., & Collins, D. L.
(2013). Automated segmentation of basal ganglia and deep brain structures in MRI of Parkinson’s disease [Comparative Study Research Support, Non-U.S.
Gov’t].International Journal of Computer Assisted Radiology and Surgery, 8(1), 99–110.http://dx.doi.org/10.1007/s11548-012-0675-8.
Hoehn, M. M., & Yahr, M. D. (1967). Parkinsonism: Onset, progression and mortality.
Neurology, 17(5), 427–442.
Huebl, J., Schoenecker, T., Siegert, S., Brucke, C., Schneider, G. H., Kupsch, A., ... Kuhn, A. A. (2011). Modulation of subthalamic alpha activity to emotional stimuli correlates with depressive symptoms in Parkinson’s disease. Movement Disorders, 26(3), 477–483.http://dx.doi.org/10.1002/mds.23515.
Hughes, A. J., Daniel, S. E., Kilford, L., & Lees, A. J. (1992). Accuracy of clinical diagnosis of idiopathic Parkinson’s disease: A clinico-pathological study of 100 cases.Journal of Neurology, Neurosurgery and Psychiatry, 55(3), 181–184.
Kotz, S. A., Meyer, M., Alter, K., Besson, M., von Cramon, D. Y., & Friederici, A. D.
(2003). On the lateralization of emotional prosody: An event-related functional MR investigation.Brain and Language, 86(3), 366–376.
Kotz, S. A., & Schwartze, M. (2010). Cortical speech processing unplugged: A timely subcortico-cortical framework. Trends in Cognitive Sciences, 14(9), 392–399.
http://dx.doi.org/10.1016/j.tics.2010.06.005.
Kühn, A. A., Hariz, M. I., Silberstein, P., Tisch, S., Kupsch, A., Schneider, G. H., ...
Brown, P. (2005). Activation of the subthalamic region during emotional processing in Parkinson disease.Neurology, 65(5), 707–713.http://dx.doi.org/
10.1212/01.wnl.0000174438.78399.bc.
Langston, J. W., Widner, H., Goetz, C. G., Brooks, D., Fahn, S., Freeman, T., & Watts, R.
(1992). Core assessment program for intracerebral transplantations (CAPIT).
Movement Disorders, 7(1), 2–13.
Leaver, A. M., & Rauschecker, J. P. (2010). Cortical representation of natural complex sounds: Effects of acoustic features and auditory object category [Research Support, N.I.H., Extramural Research Support, U.S. Gov’t, Non-P.H.S.].Journal of Neuroscience, 30(22), 7604–7612. http://dx.doi.org/10.1523/JNEUROSCI.0296- 10.2010.
Maris, E., & Oostenveld, R. (2007). Nonparametric statistical testing of EEG- and MEG-data.Journal of Neuroscience Methods, 164(1), 177–190.http://dx.doi.org/
10.1016/j.jneumeth.2007.03.024.
Mattis, S. (1988).Dementia rating scale. Odessa, FL: Psychological Assessment Resources Inc..
Morris, J. S., Scott, S. K., & Dolan, R. J. (1999). Saying it with feeling: Neural responses to emotional vocalizations.Neuropsychologia, 37(10), 1155–1163.
Nelson, H. (1976). A modified card sorting test sensitive to frontal lobe defects.
Cortex, 12, 313–324.
Oostenveld, R., Fries, P., Maris, E., & Schoffelen, J. M. (2011). FieldTrip: Open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data [Research Support, Non-U.S. Gov’t]. Computational Intelligence and Neuroscience, 2011, 156869.http://dx.doi.org/10.1155/2011/156869.
Overath, T., Kumar, S., von Kriegstein, K., & Griffiths, T. D. (2008). Encoding of spectral correlation over time in auditory cortex [Research Support, Non-U.S.
Gov’t].Journal of Neuroscience, 28(49), 13268–13273.http://dx.doi.org/10.1523/
JNEUROSCI.4596-08.2008.
Paulmann, S., & Kotz, S. A. (2008). An ERP investigation on the temporal dynamics of emotional prosody and emotional semantics in pseudo- and lexical-sentence context.Brain and Language, 105(1), 59–69.
Péron, J., Cekic, S., Haegelen, C., Sauleau, P., Patel, S., Drapier, D., ... Grandjean, D.
(2015). Sensory contribution to vocal emotion deficit in Parkinson’s disease after subthalamic stimulation.Cortex, 63, 172–183.http://dx.doi.org/10.1016/
j.cortex.2014.08.023.
Péron, J., El Tamer, S., Grandjean, D., Leray, E., Travers, D., Drapier, D., ... Millet, B.
(2011). Major depressive disorder skews the recognition of emotional prosody.
Progress in Neuro-Psychopharmacology and Biological Psychiatry, 35, 987–996.
Péron, J., Frühholz, S., Ceravolo, L., & Grandjean, D. (2016). Structural and functional connectivity of the subthalamic nucleus during vocal emotion decoding.Social Cognitive and Affective Neuroscience, 11(2), 349–356.http://dx.doi.org/10.1093/
scan/nsv118.
Péron, J., Frühholz, S., Vérin, M., & Grandjean, D. (2013). Subthalamic nucleus: A key structure for emotional component synchronization in humans.Neuroscience and Biobehavioral Reviews, 37(3), 358–373. http://dx.doi.org/10.1016/j.
neubiorev.2013.01.001.
Péron, J., Grandjean, D., Le Jeune, F., Sauleau, P., Haegelen, C., Drapier, D., ... Vérin, M.
(2010). Recognition of emotional prosody is altered after subthalamic nucleus deep brain stimulation in Parkinson’s disease. Neuropsychologia, 48(4), 1053–1062.http://dx.doi.org/10.1016/j.neuropsychologia.2009.12.003. S0028- 3932(09)00477-1 [pii].
Reitan, R. (1958). Validity of the trail making test as an indication of organic brain damage.Perceptual and Motor Skills, 8, 271–276.
Sander, D., Grandjean, D., Pourtois, G., Schwartz, S., Seghier, M. L., Scherer, K. R., &
Vuilleumier, P. (2005). Emotion and attention interactions in social cognition:
Brain regions involved in processing anger prosody.Neuroimage, 28(4), 848–858.
Schirmer, A., & Kotz, S. A. (2006). Beyond the right hemisphere: Brain mechanisms mediating vocal emotional processing.Trends in Cognitive Sciences, 10(1), 24–30.
Schwab, R., & England, A. (1969). Projection techniques for evaluating surgery in Parkinson’s disease. InPaper presented at the Third Symposium on Parkinson’s Disease, May 20–22, 1968. Royal College of Surgeons in Edinburgh.
