Functional brain changes underlying irritability in premanifest Huntington's disease

Abstract

The clinical phenotype of Huntington's disease (HD) consists of motor, cognitive and psychiatric symptoms, of which irritability is an important manifestation. Our aim was to identify the functional and structural brain changes that underlie irritability in premanifest HD (preHD). Twenty preHD carriers and 20 gene‐negative controls from HD families took part in the study. Although the 5‐year probability of disease onset was only 11%, the preHD group showed striatal atrophy and increased clinical irritability ratings. Functional MRI was performed during a mood induction experiment by means of recollection of emotional (angry, sad, and happy) and neutral autobiographical episodes. While there were no significant group differences in the subjective intensity of the emotional experience, the preHD group showed increased anger‐selective activation in a distributed network, including the pulvinar, cingulate cortex, and somatosensory association cortex, compared to gene‐negative controls. Pulvinar activation during anger experience correlated negatively with putaminal grey matter volume and positively with irritability ratings in the preHD group. In addition, the preHD group showed a decrease in anger‐selective activation in the amygdala, which correlated with putaminal and caudate grey matter volume. In conclusion, compared to gene‐negative controls, anger experience in preHD is associated with activity changes in a distributed set of regions known to be involved in emotion regulation. Increased activity is related to behavioral and volumetric measures, providing insight in the pathophysiology of early neuropsychiatric symptoms in preHD. Hum Brain Mapp 36:2681–2690, 2015. © 2015 Wiley Periodicals, Inc.

INTRODUCTION

Huntington's disease (HD) is a dominantly inherited, fatal neurodegenerative disorder caused by a CAG repeat expansion in the HTT gene on chromosome 4. In the western world, the prevalence of HD is 5–10 per 100,000, making it the third most prevalent neurodegenerative disease after Alzheimer's disease and Parkinson's disease. The principal neuropathological change is the loss of striatal medium spiny neurons, and to a lesser degree neuronal loss in neocortex, hippocampus, and thalamus. Progression of the disease is associated with more widespread neuronal loss and brain atrophy [Vonsattel, 2008]. HD is typically characterized by a triad of symptoms and signs, composed of a movement disorder including chorea, cognitive deterioration, and behavioral disturbances. Irritability is among the earliest and most frequent psychiatric symptoms and is associated with inflexibility, perseverative preoccupations, verbal outbursts, and aggression [Craufurd et al., 2001]. It strongly contributes to the need for inpatient treatment [Wheelock et al., 2003]. Clinical disease onset is conventionally defined by the onset of the movement disorder. However, neuropsychiatric symptoms are often already present before the onset of motor symptoms. Carriers of the Huntington mutation that do not display motor disturbance are referred to as premanifest HD (preHD) subjects [Ross et al., 2014]. Recently, research has increasingly focused on the premanifest stage, as this might benefit the development of therapeutic interventions [Tabrizi et al., 2013]. Irritability has been reported in up to 73% of preHD subjects [van Duijn et al., 2007].

Little is known about the brain changes that underlie neuropsychiatric symptoms in preHD. One study examined irritability in preHD by giving negative false feedback in a size discrimination task. The results showed reduced activation of the amygdala as well as reduced coupling between amygdala and medial prefrontal cortex in the preHD group compared to healthy controls [Kloppel et al., 2010]. This may reflect an abnormal recognition of the own emotional state in preHD, a finding that has been reported in manifest HD [Ille et al., 2011; Reedeker et al., 2012] and may also underlie impaired recognition of emotions in others [de Gelder et al., 2008; Henley et al., 2012; Ille et al., 2011; Kret et al., 2010; Trinkler et al., 2013]. In this study, we build on these findings and focus on the experiential aspect of emotions, that is, “feelings” [Damasio and Carvalho, 2013].

The main goal of this study was to examine the pathophysiology underlying irritability in preHD. We, therefore, measured brain activation associated with feelings of anger in preHD and gene‐negative controls from HD families (Gene‐Neg Ctrls). The neural circuitry underlying anger experience in normal subjects comprises a distributed set of cortical and subcortical regions, of which the pulvinar, cingulate cortex, lenticular nucleus, insula, frontal cortices, cerebellum, and (para)hippocampal regions have been consistently reported [Damasio and Carvalho, 2013; Denson et al., 2009; Fabiansson et al., 2012; Pawliczek et al., 2013; Vytal and Hamann, 2010]. It has been proposed that areas containing interoceptive topographic maps like the insula underlie the subjective experience of emotional bodily states (feelings), while the anterior cingulate cortex (ACC) generates emotion‐related actions [Damasio and Carvalho, 2013]. Interestingly, thalamic, cerebellar, and cingular metabolic abnormalities have also been reported in preHD [Feigin et al., 2007]. We investigated group differences in the functional neuroanatomy of experiencing anger and related this to structural brain changes and clinical irritability.

MATERIALS AND METHODS

Participants

The study was approved by the Ethical Committee of University Hospitals Leuven (ML8040) and subjects’ consent was obtained according to the Declaration of Helsinki. All participants were recruited through the department of Human Genetics of University Hospitals Leuven following predictive genetic testing for HD. After complete description of the study to the subjects, written informed consent was obtained. Twenty premanifest carriers of the HD mutation (CAG repeat length of 40 or more; absence of visible chorea and Unified Huntington's Disease Rating Scale (UHDRS) motor score <6) [Tabrizi et al., 2009] participated as well as twenty gene‐negative controls (Gene‐Neg Ctrls) from HD families (CAG repeat length of 35 or less). We included Gene‐Neg Ctrls as a control group to match both groups regarding the distress of growing up with affected and at‐risk family members and undergoing predictive genetic testing for HD [Julien et al., 2007]. Demographic data are summarized in Table 1. All participants underwent motor assessment based on the motor subscale of the Unified Huntington's Disease Rating Scale [Huntington Study Group, 1996], administered by a trained physician. Cognitive evaluation consisted of a battery of commonly used neuropsychological tests. Mood evaluation included BDI [Beck et al., 1961] and STAI [Spielberger et al., 1970] (Table 1). To assess neuropsychiatric symptoms associated with HD, the Problem Behaviors Assessment for HD (PBA‐HD) [Craufurd et al., 2001] was administered to all participants by a trained physician who was unaware of the genetic status of the interviewee. The PBA‐HD consists of a 36‐item semistructured interview, with a three factor structure: Apathy, Irritability, and Depression.

Table 1.

Demographic, clinical, behavioral, and cognitive characteristics of premanifest Huntington's disease carriers and gene‐negative controls

	Premanifest carriers of Huntington mutation (N=20)		Gene‐Neg Ctrls (N=20)
Characteristic	N	%	N	%	P
Gender (female)	13	65	10	50	0.52a
	Mean	SD	Mean	SD
Age (years)	36.7	8.0	38.4	14.4	0.64b
CAG	42.1	1.8	20.1	4.6	0.0010b*
UHDRS_m	2.7	1.9	1.8	1.2	0.10b
5yPDO	0.11	0.16	0	0
Mood
BDI	5.8	6.1	4.2	5.4	0.50b
STAI 1	43.4	9.3	44.5	4.6	0.58b
STAI 2	45.1	7.2	47.5	4.2	0.21b
PBA‐HD Total	11.2	11.1	4.7	4.3	0.023b*
PBA‐HD Apathy	2.5	3.6	0.5	0.7	0.024b*
PBA‐HD Irritability	1.85	2.35	0.65	0.67	0.039b*
PBA‐HD Depression	2.35	2.9	1.15	1.6	0.11b
Cognition
MMSE	29.0	1.3	29.3	1.2	0.52b
RAVLT A1‐A5	52.4	9.2	53.3	9.1	0.76b
RAVLT A7	10.7	2.39	12.15	2.18	0.052b
RAVLT A8	13.9	1.23	14.55	1.15	0.12b
Stroop1	44.7	8.2	44.1	7.1	0.81b
Stroop2	58.2	12.8	56.4	9.83	0.63b
Stroop3	86.9	17.4	86.7	17.3	0.98b
DS forward	6.5	0.9	6.4	1.3	0.67b
DS backward	4.6	1.3	5.0	1.2	0.26b
BNT	54.8	3.2	54.3	3.5	0.64b
TMT A	29.0	1.3	29.3	1.2	0.66b
TMT B	67.6	22.7	59.8	18.6	0.25b
AVF	23.7	5.4	23.6	4.3	0.92b
LVF	31.6	13.5	38.5	10.1	0.074b
NART	86.8	5.0	91.4	5.0	0.15b
Digit Symbol	79.5	24.7	81.3	17.5	0.79b
VAS
Anger	61.5	16.8	72.6	18.9	0.058b
Sad	67.4	19.0	78.5	18.3	0.067b
Happy	75.5	15.1	76.2	13.4	0.89b
Neutral	74.4	16.9	74.4	15.9	0.99b

CAG, number of CAG‐repeats; UHDRS_m, Unified Huntington's Disease Rating Scale motor score; 5yPDO, 5 year probability of disease onset; BDI, Beck Depression Inventory; STAI, State‐Trait Anxiety Inventory; PBA‐HD: Problem Behaviours Assessment for Huntington's Disease; MMSE, Mini Mental‐State Examination; RAVLT, Rey Auditory Verbal Learning Test; A1–A5 = the sum of scores on trials A1 to A5 of the RAVLT; A7: Delayed Recall score; A8: Recognition score (difference between the number of correct hits and false hits on the recognition trial); DS: digitspan; BNT, Boston Naming Test; TMT, Trail Making Test; AVF, Animal Verbal Fluency (# in 1 minute); LVF, letter verbal fluency with # in 1‐minute starting with “N,” “A,” and “K”; NART: National Adult Reading Test (Dutch version); Digit Symbol: Digit Symbol subtest from Wechsler Adult Intelligence Scale R, VAS, Visual Analogous Scale judgments of intensity of emotion re‐experience (0: very low intensity, 100: very high intensity).

^aBased on 2‐sample test for equality of proportions with continuity correction.

^bBased on 2‐sample t‐test.

*Significant difference

Experimental Stimuli and Paradigm

One week prior to the MRI session, participants were instructed to select eight autobiographical events that were associated with intense emotions: two during which they felt angry, two during which they felt happy, two during which they felt sad and finally two events that were not associated with any particular emotion. In line with a previous study, there were no a priori constraints of any kind to the themes or nature of the events [Damasio et al., 2000]. In addition, participants were instructed to provide a word for each of the eight events so that when the word was shown to them, they would immediately recognize to which event and emotion it referred.

The experiment consisted of two runs, each containing four blocks, one for every emotion. The procedure of one block is schematically presented in Figure 1. A block started with a 1,000 ms black screen, followed by a 3,000 ms presentation of one of the words provided by the participant. Next, “Close your eyes now” appeared for 2,000 ms. Subsequently, a blank screen was shown for 61 s. This 61 s event was the emotion experience event. The end of this interval was signaled to the participant by three alternating 500 ms presentations of a black and white screen (which were clearly noticeable even with the eyes closed). During the following 10 s, a visual analogous scale was presented on which participants could indicate how intensely they had re‐experienced the emotion. Next, the text “Press after vertical line” was presented for 4,000 ms. Subsequently, 30 stimuli consisting of a circle filled with line gratings were serially presented for 500 ms, with a 500 ms interstimulus interval during which a black screen was presented. These 30 grating stimuli consisted of 5 vertical and 25 horizontal gratings. These grating orientation detection tasks alternated with the emotion experience events to minimize carry‐over effects of emotions between emotion experience events.

Figure 1.

Schematic presentation of the experimental procedure.

A practice session immediately preceded the scanning session. The procedure of the practice session was identical to the experimental procedure, except for the duration of the emotion experience event, which lasted 30 instead of 61 s. The participant was instructed to re‐experience the respective emotion as intensely as possible during the emotion experience event.

The method of emotion induction through recollection of autobiographical episodes has been proven valuable [Damasio et al., 2000]. We reasoned that it might be beneficial to minimize contamination of subjective experience of different emotions by introducing a clear psychological demarcation episode between emotions. Therefore, we added a nonemotional perceptual task that showed considerable cognitive (attentional) demands as a means to obtain a form of “tabula rasa” of the subjective emotional state. Furthermore, for standardization and follow‐up (the postscanning interview, see below) purposes, we included the word selection by the participant that was visually presented in the scanner.

Image Acquisition

Functional MR imaging was performed using a 3 Tesla MR scanner (Achieva3T; Philips, Best, the Netherlands) with a 32‐channel head coil. All participants underwent two runs, each with a duration of 507 s. In each run 140 T2*‐weighted Blood‐Oxygenation Level Dependent (BOLD) contrast volumes were acquired. A volume consisted of 70 axial slices oriented parallel to the anterior commissure‐ posterior commissure (AC‐PC) plane (slice thickness=2.0 mm; no gap; in‐plane resolution= 2.75 × 2.75 mm; matrix size=80 × 80; FOV=220 × 220 mm), positioned to cover the whole brain (TE=26; TR=3,500 ms; flip angle=90°), optimized for subcortical sensitivity [Morawetz et al., 2008]. Functional runs were preceded by acquisition of four dummy volumes to allow for T1 equilibration. A high‐resolution T1‐weighted anatomical image (voxel size=1 × 1×1 mm) was acquired in between the functional runs, using a three‐dimensional (3D) magnetization‐prepared‐rapid acquisition gradient echo sequence (TR = 9.6 ms; TE = 4.6 ms; matrix size=256 × 256; 182 slices).

Post‐Scanning Semistructured Interview

Following the scanning session, all participants were interviewed by a physician who was blind regarding the genetic status. The semistructured interview consisted of standardized questions related to the words (and hence the autobiographical emotional episodes) selected by the participant and inquired about the content and chronology of the episode as well as about the re‐experience in the scanner. In addition to the subjective ratings during the scanning session, these interviews confirmed that the participants successfully re‐experienced the target emotions.

Analysis

Grey matter volume analyses were performed by means of Voxel‐based morphometry using the VBM8 toolbox (http://dbm.neuro.uni-jena.de/vbm) for SPM8 (Wellcome Trust Centre for Neuroimaging, UCL, London, United Kingdom) within MatLab R2008a (Mathworks, Inc). VBM has been reported as a valuable tool to assess subcortical grey matter volume in HD [Douaud et al., 2006]. Preprocessing of T1 structural images included bias correction, segmentation and normalization to MNI space within a unified model, including high‐dimensional DARTEL‐normalization. Grey matter was modulated by affine and nonlinear components. Modulated images were smoothed using a Gaussian kernel of 8 mm at FWHM.

Other imaging data were analyzed using BrainVoyager QX 2.8.2 (Brain Innovation, Maastricht, The Netherlands). Preprocessing of the functional data included mean intensity adjustment, slice scan time correction (cubic‐spline interpolation), 3D motion correction (trilinear/sinc interpolation), temporal filtering (high pass GLM‐Fourier of 2 sines/cosines) and Gaussian spatial smoothing (8 mm). Functional data were coregistered with the anatomical volume and transferred into Talairach (TAL) space. Analysis of cortical activation followed cortex‐based intersubject alignment, which minimized interindividual differences in cortical folding pattern by maximizing the spatial correspondence between the individual gyral/sulcal patterns. Noncortical effects were analyzed in subcortical space as defined by a commonly used atlas [Eickhoff et al., 2007] that was adapted for TAL‐space. The statistical analysis was based on the General Linear Model in an extended TAL‐space, including the cerebellum. Emotion blocks were defined by the 61 s time interval synchronized with the on‐ and offset of the blank screen. We performed a parametric single‐trial effect coding at subject level in which every emotion block was parametrically modulated by the intensity rating following the respective event. Subsequently, we performed a random effects analysis with the four emotion events and 4 visual reaction time events as within‐subject variables and HD genetic status (preHD, Gene‐Neg Ctrls) as between subjects factor. The statistical threshold was set at 0.05 (cluster‐size corrected following 5,000 Monte‐Carlo simulations). Demographic and psychophysical data were analyzed with IBM SPSS Statistics 20 and R.

RESULTS

Behavioral Results

Table 1 shows the demographic, clinical, motor, behavioral and cognitive characteristics of the two study groups. Significant group differences were restricted to the CAG‐repeat length and PBA‐HD total score as well as the PBA‐HD subscales apathy and irritability. Probability of disease onset within 5 years (5yPDO) was estimated based on age and CAG repeat length [Langbehn et al., 2004]. This averaged only 11%, indicating that the preHD group was far from a manifest stage. There were no significant group differences in the subjective intensity ratings of emotional experiences during the scanning session, although there was a trend (P =0.058) toward less intense subjective anger experience in the preHD group.

Structural Imaging Results

Two bilateral clusters in the striatum showed reduced grey matter volume in the preHD group (P<0.001, uncorrected; >70 voxels). These clusters were located in the head of the caudate nucleus and putamen (See Fig. 2). We pooled the left caudate nucleus cluster with the right caudate cluster and the left putamen cluster with the right. We then calculated the amount of grey matter in the caudate nucleus and putamen clusters for every preHD subject and the Pearson correlation coefficient with the 5yPDO and PBA‐HD total and subscale scores. This revealed for both caudate nucleus and putamen significant negative correlations with 5yPDO (caudate: r(20)=−0.600, P<0.005; putamen r(20)=−0.678, P<0.001), but no significant correlations with any of the PBA‐HD scores.

Figure 2.

Structural imaging results. The top row displays the regional subcortical volumetric group differences. The scatterplots below show the correlation with 5‐year probability of disease onset. The solid line displays the linear regression fit and the dotted lines indicate the mean 95% confidence band. The x‐axis scale refers to the amount of grey matter within the region demarking the group difference.

Functional Imaging Results

We investigated group differences in anger‐selective (anger experience > sad+happy+neutral experience) brain activation between the preHD and Gene‐Neg Ctrls group. The results are presented in Figure 3 and Table 2. We observed hyperactivation in the preHD group in the right pulvinar, premotor/prefrontal cortex, parieto‐occipital sulcus, left pallidum, cerebellum, supramarginal gyrus, superior parietal lobule (SPL), and bilateral anterior and posterior cingulate cortex, while we observed hypoactivation in the left amygdala.

Figure 3.

Functional imaging results. The two top rows display the statistical groups differences (preHD>Gene‐Neg Ctrls) during experience of anger emotions (compared to happy, sad and neutral experiences). The third row displays scatterplots of within preHD‐group associations between the fMRI parameter estimates and volumetric and behavioral data. The solid line displays the linear regression fit and the dotted lines indicate the mean 95% confidence band. The x‐axis scale refers to the number of grey matter voxels within the region demarking the group difference. The bottom row displays the statistical groups difference (Gene‐Neg Ctrls>preHD) during experience of anger emotions (compared to happy, sad and neutral experiences) and scatterplots of the association with volumetric data.

Table 2.

Structural and functional imaging results

		Peak TAL coordinate
	N voxels (vertices)	x	y	z	T	P
Structural group differences
(Gene‐Neg Ctrls>preHD)
Right caudate nucleus	93	19	18	1	3.24	0.001
Left caudate nucleus	309	−8	20	0	4.16	0.0002
Right putamen	386	28	−12	−1	3.71	0.0007
Left putamen	378	−28	−15	−4	3.50	0.001
(preHD>Gene‐Neg Ctrls)
No significant results
Functional group differences
(preHD>Gene‐Neg Ctrls)
Right Thalamus (pulvinar)	1,133	23	−29	6	3.33	0.009
Left lenticular nucleus	431	−16	−8	6	2.81	0.007
Left cerebellum	1,390	−43	−44	−42	3.21	0.002
Right cingulate cortex
Anterior	(511)	11	20	36	3.65	0.0007
Posterior	(670)	4	−30	30	3.86	0.0004
Left cingulate cortex
Anterior	(214)	−5	16	29	4.34	0.0001
Posterior	(412)	−4	−33	29	4.80	0.00002
Right POS	(282)	22	−57	16	3.73	0.0006
Right SFS	(338)	2	11	23	3.44	0.001
Left STG	(698)	−56	−34	14	3.95	0.0003
Left SPL	(568)	−30	−49	38	2.92	0.005
(Gene‐Neg Ctrls>preHD)
Left amygdala	458	−16	−2	−21	3.88	0.0003

Gene‐Neg Ctrls, Gene‐negative control group; preHD, premanifest carriers of Huntington's Disease group; POS, parieto‐occipital sulcus; SFS, superior frontal sulcus; STG, superior temporal gyrus; SPL, superior parietal lobule.

Correlations with Structural Results

Next, we investigated whether anger‐selective regional activation changes in the preHD group were associated with striatal atrophy. Therefore, we correlated the individual grey matter volumes of the bilateral head of the caudate nucleus and putamen with the activation magnitude during anger experience. Correlations with peak values are reported. For caudate volume, there was a positive correlation with activation in the left posterior cingulum (r(20)=0.721, P<0.0003) and amygdala (r(20)=0.506, P<0.023). For putaminal grey matter volume, there was a positive correlation with amygdalar activation (r(20)=0.645, P<0.002) and a negative correlation with pulvinar activation (r(20)=−0.546, P<0.013) (Fig. 3).

Brain‐Behavior Correlations

The association between clinical irritability and anger experience was examined by correlating the PBA‐HD Irritability scores with the functional MRI (fMRI) activations during anger experience (vs. all other emotions). The results show a positive correlation with pulvinar activation (r(20)=0.680, P<0.001) (Fig. 3).

DISCUSSION

The main objective was to investigate the neural pathophysiology underlying irritability in carriers of the Huntington mutation before the onset of motor symptoms. The estimated probability of disease onset within 5 years in these preHD subjects averaged only 11%. Nevertheless, the preHD group showed striatal atrophy and increased irritability compared to a control group that faced similar HD‐related familial stressors. This result is consistent with previous reports of early presymptomatic structural brain changes and psychiatric features in HD [Aylward et al., 2004; Tabrizi et al., 2013; van Duijn et al., 2007].

The main finding of this study concerns the neural changes in preHD associated with the experience of feelings of anger. The results reveal functional group differences in a distributed set of areas that have been associated with emotion regulation in normal subjects: lateral cerebellum, pallidum, anterior and posterior cingulate cortex, somatosensory association cortex, pulvinar, and amygdala [Damasio et al., 2000; Denson et al., 2009; Fabiansson et al., 2012; Pawliczek et al., 2013; Vytal and Hamann, 2010]. Mainly, we found that feelings of anger, relative to other emotions and neutral state, in preHD subjects are associated with hyperactivation in the emotion experience neurocircuitry.

Interestingly, the hyperactivation in the ACC falls within the rostral cingulate zone (RCZ). This area forms part of the basal ganglia thalamocortical circuit [Alexander et al., 1986] and receives dopaminergic projections from the substantia nigra. It projects to the dorsal striatum, motor cortices and amygdala [Shackman et al., 2011]. All these areas show functional or structural abnormalities in this study. The RCZ is a somatotopically organized premotor area that has been associated with the expression of affect. For instance, direct microstimulation of the monkey RCZ triggers emotional expressions [Shackman et al., 2011]. The human RCZ has also been associated with the experience of affect [Lane et al., 1998] and cognitive control [Shackman et al., 2011]. Remarkably, emotional expressions, experience of affect, and cognitive control are all affected in HD, suggesting a central role of ACC abnormalities in the irritability phenotype.

In addition to the ACC, the thalamic pulvinar seems to play a cardinal role in irritability symptomatology. The pulvinar is also embedded in the basal ganglia thalamocortical circuit [Alexander et al., 1986] and has functionally been associated (besides vision) with emotional experience, particularly feelings of anger [Damasio et al., 2000; Denson et al., 2009; Vytal and Hamann, 2010]. It is noteworthy that all these associations are reflected in the present findings. First, pulvinar activation during anger experience correlated negatively with grey matter volume in the atrophic putamen in the preHD group. Hence, more atrophy at the striatal level of the basal ganglia thalamocortical circuit is accompanied by higher activation at the thalamic level. Second, the pulvinar was more active in the preHD than in the Gene‐Neg Ctrls during anger experience. This represents pathology in a node of the normal network of experiencing anger [Damasio et al., 2000]. Third, pulvinar activation correlated with PBA‐HD Irritability scores. This is consistent with functional imaging and lesion studies documenting the involvement of the pulvinar in irritability and emotion dysregulation in other neuropsychiatric conditions [Deveney et al., 2013; Dougherty et al., 2004; Lee et al., 2014; Leibenluft, 2011; Liebermann et al., 2013].

Although the earliest structural macroscopic pathology in HD is observed in the striatum, there is evidence that the thalamus and ACC also show structural and functional neuropathologic changes in the course of HD. Atrophy of the ACC has been documented at the early manifest and even premanifest stage [Hobbs et al., 2011]. Furthermore, the ACC volume correlated with emotion processing measures (facial emotion recognition) and mood, that is, self‐reported depression (Beck Depression Inventory score) [Hobbs et al., 2011]. Functional abnormalities of the ACC in preHD have been related to cognitive deficits, that is, response inhibition [Rao et al., 2014]. Regarding the thalamus, there is evidence suggesting an increase in thalamic metabolism at the premanifest stage, followed by a decrease at the transition to the manifest disease stage [Eidelberg and Surmeier, 2011; Feigin et al., 2007; Tang et al., 2013]. Possibly, the initial hyperactivity reflects a compensatory mechanism. The present results are in line with this notion and provide a link with the functional neuroanatomy of anger regulation and neuropsychiatric symptomatology.

In addition, we observed increased activation in the somatosensory cortex in the preHD group. The hyperactive region falls within the somatosensory association cortex in an area that has been associated with awareness of emotions and internal body states [Anders et al., 2004; Damasio, et al., 2000]. This result is thus more compatible with the notion that preHD subjects experience abnormal bodily sensations (feelings) during anger, regardless of subjective intensity of the experience. However, it may be that an abnormal experience of bodily sensations hampers their recognition as emotional states [Kloppel et al., 2010]. This notion fits with the reduced activation of the amygdala. The present results also indicate that this amygdalar effect is proportional to both putaminal and caudate grey matter volume.

Taken together, the present evidence indicates that the neural pathophysiology underlying irritability in preHD constitutes a hyperactivation of the neural anger circuitry, in which the ACC and thalamic pulvinar play a central role.