NeuroImage: Clinical 21 (2019) 101662 Contents lists available at ScienceDirect NeuroImage: Clinical journal homepage: www.elsevier.com/locate/ynicl Lack of response to disgusting food in the hypothalamus and related T structures in Prader Willi syndrome ⁎ Laura Blanco-Hinojoa,b, Jesus Pujola,b, , Susanna Esteba-Castilloc, Gerard Martínez-Vilavellaa, Olga Giménez-Palopd, Elisabeth Gabaue, Laia Casamitjanad, Joan Deusa,f, Ramón Novellc, Assumpta Caixàsd a MRI Research Unit, Department of Radiology, Hospital del Mar, 08003 Barcelona, Spain b Centro Investigación Biomédica en Red de Salud Mental, CIBERSAM G21, 08003 Barcelona, Spain c Specialized Service in Mental Health and Intellectual Disability, Institut Assistència Sanitària (IAS), Parc Hospitalari Martí i Julià, 17190 Girona, Spain d Endocrinology and Nutrition Department, Parc Taulí Hospital Universitari, Institut d'Investigació i Innovació Parc Taulí I3PT- UAB, 08208 Sabadell, Spain e Clinical Genetics, Pediatrics Department, Parc Taulí Hospital Universitari, Institut d'Investigació i Innovació Parc Taulí I3PT- UAB, 08208 Sabadell, Spain f Department of Clinical and Health Psychology, Autonomous University of Barcelona, 08193 Barcelona, Spain A R T I C LE I N FO A B S T R A C T Keywords: Objective: To investigate, based on a putative abnormal neural processing of disgusting signals in Prader Willi Prader Willi syndrome syndrome (PWS) patients, the brain response to visual representations of disgusting food in PWS using functional Functional MRI MRI (fMRI). Disgust Methods: Twenty-one genetically-confirmed PWS patients, 30 age- and sex-matched and 28 BMI-matched con- Hypothalamus trol subjects viewed a movie depicting disgusting food-related scenes interspersed with scenes of appetizing food while fMRI was acquired. Brain activation maps were compared between groups and correlated with disgust and hunger ratings. Results: At the cortical level, the response to disgusting food representations in PWS patients was qualitatively similar to that of control subjects, albeit less extensive, and engaged brain regions typically related to visually- evoked disgust, such as the anterior insula/frontal operculum, the lateral frontal cortex and visual areas. By contrast, activation was almost absent in limbic structures directly concerned with the regulation of instinctive behavior robustly activated in control subjects, such as the hypothalamus, amygdala/hippocampus and peria- queductal gray. Conclusions: Our study provides novel insights into the neural substrates of appetite control in a genetically- mediated cause of obesity. The presence of significant cortical changes further indicates that PWS patients consciously process disgusting stimuli, but the virtual absence of response in deep, limbic structures suggests that disgusting signals do not adequately reach the primary brain system for the appetite control. 1. Introduction impaired satiety mechanisms (Hinton et al., 2006; Holland et al., 1995; Shapira et al., 2005) secondary to hypothalamic dysfunction Prader Willi syndrome (PWS) is a genetic disorder caused by the (Goldstone, 2004, 2006; Swaab, 1997; Swaab et al., 1995) and related lack of expression of paternal genes in the chromosome 15, resulting in structures also implicated in the regulation of feeding (Dimitropoulos a complex phenotype that presents with characteristic physical traits and Schultz, 2008; Hinton et al., 2006; Holsen et al., 2006; Pujol et al., and specific endocrine and behavioral problems (Cassidy et al., 2012). 2016; Zhang et al., 2013). Altered control of appetite is generally a major issue for PWS patients in Although PWS patients appear to show an adequate general the form of a strong desire for food and a tendency to overeat with knowledge about food (Dykens, 2000), indiscriminate eating is a subsequent obesity (Cassidy et al., 2012; McAllister et al., 2011; Miller common phenomenon (Dykens, 2000; Dykens et al., 2007). Specific et al., 2011; Tauber et al., 2014). Although the origin of hyperphagia in eating behaviors include consuming products commonly considered as PWS remains incompletely understood, it has been associated with unappealing or inedible, food from the trash and decayed food (Dykens ⁎ Corresponding author at: MRI Department, Hospital del Mar, Passeig Marítim 25-29, 08003 Barcelona, Spain. E-mail addresses: 21404jpn@comb.cat (J. Pujol), susanna.esteba@ias.cat (S. Esteba-Castillo), egabau@tauli.cat (E. Gabau), lcasamitjana@tauli.cat (L. Casamitjana), joan.deus@uab.cat (J. Deus), ramon.novell@ias.scs.es (R. Novell). https://doi.org/10.1016/j.nicl.2019.101662 Received 18 April 2018; Received in revised form 22 November 2018; Accepted 4 January 2019 Available online 04 January 2019 2213-1582/ © 2019 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/). L. Blanco-Hinojo et al. NeuroImage: Clinical 21 (2019) 101662 et al., 2007; Russell and Oliver, 2003), which suggests the hypothesis Helsinki. Written informed consent was obtained from caregivers and that the altered control of appetite could partly result from the in- control subjects. Verbal or written assent was also obtained from the adequate responding to disgust-evoking stimuli in brain systems reg- PWS patients. ulating feeding. Besides the potential health risk of eating spoiled food (Chapman and Anderson, 2012), a deficient processing of disgusting 2.2. Stimulus and experimental procedure signals could contribute to overeating and obesity in PWS via a higher threshold for rejecting food products (Houben and Havermans, 2012). A 6-m color movie was used involving the successive presentation of We used functional MRI (fMRI) to examine the brain response to appetizing food-related scenes as baseline condition and 8 scenes of visual representations of disgusting food in a group of PWS patients disgusting food interspersed throughout the movie at pseudorandom compared with a healthy control group. Participants watched a movie intervals (Pujol et al., 2018). Each disgusting scene lasted 8–10 s. In the showing disgusting food scenes interspersed with baseline scenes of appetizing condition, the scenes presented palatable food items in res- appetizing food. Brain activation maps were compared between both taurants and gastronomy scenes, whereas the disgust-evoking scenes groups and correlated with disgust and hunger ratings. showed food-related disgust elicitors such as decaying food, maggots and cockroaches in food and people eating live worms. 2. Material and methods Participants underwent fMRI approximately 60 min after receiving a standardized meal to maximally control for the experimental context. 2.1. Participants During image acquisition participants were instructed to lay still and passively view the movie displayed using MRI-compatible high re- Thirty adult PWS patients were recruited from PWS referral centers solution goggles (VisuaStim Digital System, Resonance Technology Inc., in Barcelona and Girona. Genetic testing to confirm the chromosome 15 Northridge, CA, USA). Immediately after scanning, participants were anomaly was repeated in all patients at the time of inclusion by pro- asked to rate the intensity (on a 0–100 scale) of the elicited feeling of cedures fully described elsewhere (Pujol et al., 2016). Patients younger disgust experienced during the fMRI experiment. Participants also in- than 18 years, those who had nonstable medical conditions and those dicated their subjective feeling of hunger immediately before entering considered unable to follow MRI instructions were not eligible to par- the MRI room. ticipate in the study. Nine patients were excluded on the basis of ex- cessive head motion during fMRI (7 cases) or insufficient imaging data 2.3. MRI acquisition collection (2 cases) resulting in a final sample of 21 patients. A total of 9 patients were taking psychiatric medication (fluoxetine, topiramate, or Brain images were acquired on a 1.5-T Signa Excite system (General both, occasionally combined with antipsychotics [n = 4]). Treated pa- Electric, Milwaukee, WI, USA) with an eight-channel phased-array head tients were on a stable medication regime for at least 3 months prior to coil and single-shot echo planar imaging (EPI) software. Functional imaging assessment. images were acquired with gradient recalled acquisition in the steady Thirty healthy subjects matched by age and sex to the PWS group state (time of repetition [TR] = 2000 ms, time of echo [TE] = 50 ms, made up the control sample. Exclusion criteria included relevant pulse angle 90°, field of view 24 cm, 64 × 64 matrix, slice thickness of medical or neurologic disorders, substance abuse or psychiatric disease 4 mm, 1.5 mm gap). Twenty-two interleaved slices were prescribed and current medical treatment or eating disorders (Pujol et al., 2018). parallel to the anterior-posterior commissure line covering the whole To control for the confounding effect of obesity, we included an addi- brain. A 6-min scan was acquired for each participant, generating 180 tional control group of thirty subjects matched by age, sex and BMI to whole brain EPI volumes. the PWS group. Two controls were excluded from the analysis due to excessive head motion during fMRI. Table 1 provides characteristics of 2.4. Image preprocessing study participants. The study was approved by the Clinical Research Ethical Imaging data were preprocessed and analyzed using the Statistical Committee-Corporació Sanitària Parc Taulí of Sabadell, Barcelona, and Parametric Mapping software package (v8) implemented in Matlab conducted according to the principles expressed in the Declaration of (The MathWorks Inc., Natick, MA). Functional images were realigned Table 1 Demographic and clinical details of the sample. PWS (n = 21) Normal-weight controls (n = 30) Obese controls (n = 28) Age, yr 27.8 ± 8.3 (18–47) 27.9 ± 7.8 (19–45) 28.8 ± 7.2 (19–46) Sex, M/F 12/9 15/15 15/13 Body mass index, kg/m2 30.5 ± 6.3 (21.5–43.4) 22.1 ± 2.0 (17.7–25) 33.1 ± 5.7 (26.1–44.4) IQ K-BIT - Total score 68.7 ± 11.4 (48–92) 111.0 ± 11.2 (90–131)a 110.9 ± 10.4 (91–137)b Verbal 76.3 ± 11.8 (50–99) 111.3 ± 12.3 (89–133) 110.9 ± 11.9 (84–133) Performance 72.7 ± 13.2 (44–102) 112.4 ± 8.4 (93–127) 112.2 ± 10.3 (89–136) Genetic subtype, no. Type I deletion 5 Type II deletion 8 Uniparental disomy 5 Imprinting defect 3 Hyperphagia Questionnairec Total score (range 0–44) 18.6 ± 9.5 (4–38) 0 ± 0 (0–0) 1.9 ± 1.7 (0–7) Behavior (range 0–20) 9.1 ± 5.2 (2−20) 0 ± 0 (0–0) 0.6 ± 1.0 (0–4) Drive (range 0–16) 6.9 ± 3.7 (1–14) 0 ± 0 (0–0) 0.8 ± 1.1 (0–3) Severity (range 0–8) 2.6 ± 2.7 (0–8) 0 ± 0 (0–0) 0.5 ± 0.9 (0–3) Values are expressed as group: mean ± standard deviation (range). PWS, Prader Willi syndrome; IQ - KBIT, Intelligence Quotient - Kaufman Brief Intelligence Test. a n = 28. b n = 27. c Dykens et al., 2007. 2 L. Blanco-Hinojo et al. NeuroImage: Clinical 21 (2019) 101662 (motion corrected) to the first volume using conventional procedures, analyses were performed in the PWS group to map the association be- spatially normalized to the standard SPM-EPI template and resliced to tween self-report ratings of disgust and hunger (separately) and acti- 2-mm isotropic resolution in Montreal Neurological Institute (MNI) vation across the whole brain. The Kbit total score was included as a space. Smoothing was executed with an 8 mm full-width at half-max- covariate in the correlation analyses as there was a trend for patients imum (FWHM) Gaussian filter. All image sequences were inspected for with lower IQ to report higher disgust ratings (r = −0.44, p = .064). potential acquisition and normalization artifacts. Data from a region showing significant correlation with self-report ratings was extracted to generate a plot illustrating the results. 2.5. Control of potential head motion effects 2.6.1. Temporal analysis of the brain response to disgust The time series were aligned to the first image volume in each To illustrate the temporal evolution of the brain response to visual participant. Motion-related regressors and estimates of global brain representations of disgusting food in PWS patients, an analysis of the signal fluctuations were included as confounding variables in our activation dynamics was conducted using procedures fully described in single-subject analyses. Within-subject, censoring-based MRI signal ar- a previous study (López-Solà et al., 2010). Briefly, the finite impulse tifact removal (scrubbing) (Power et al., 2014) was used to discard response (FIR) analysis approach (Dale and Buckner, 1997) was em- motion-affected volumes. For each participant, a motion summary ployed to obtain 12 activation maps, with a temporal resolution of 2 s measurement that combined translations and rotations was computed (1 scan), covering the activation cycle and starting from the first scan in mm (Pujol et al., 2014). Average inter-frame motion measurements after stimulus onset (time 0) to a total of 24 s (12 consecutive scans). (head position variations of each brain volume as compared to the The model involved a total implicit baseline of 176 s (88 scans). For previous volume) served as an index of data quality to flag volumes of each subject, contrast images were calculated for the 12 regressors that suspect quality across the fMRI run. We excluded participants expressed the relative BOLD signal change from baseline throughout showing > 45 (25%) image volumes with inter-frame motion > 0.2 the activation cycle. The contrast images were then entered in 12 group mm. Using this criterion, a mean of 15.9 (8.8%) volumes from the total analyses (one-sample t-tests) to generate whole-brain activation t-sta- volumes that are included in the fMRI sequence were removed in pa- tistic maps for each scan. To graphically represent the group's time tients, and a mean of 2.7 (1.5%) volumes and 11.4 (6.3%) volumes series response for the main significant brain areas, we plotted activa- were removed in healthy and BMI-matched controls, respectively. Re- tion measurements (t-values obtained from the region coordinate maining potential motion effects were controlled by including a motion showing peak activation across the cycle) against the 12 time points summary measurement for each participant as a covariate in the group (scans). analyses (Pujol et al., 2014). Lastly, a high-pass filter was used to re- move low-frequency noise (cut off period = 1/128 Hz). 2.6.2. Thresholding criteria To correct our results for multiple comparisons, non-parametric 2.6. fMRI data analysis permutation testing was conducted using standard SnPM procedures. For all statistical models employed, we used cluster-wise inference with In the single-subject (first) level analysis, an SPM contrast map for family-wise error (FWE) rate correction of PFWE < 0.05. Input para- the main effect of task (disgust vs. appetizing condition) was generated meters to SnPM included a cluster-forming threshold of p < .005 and for each participant. For this analysis, the response at each voxel was 10,000 random permutations. No variance smoothing was carried out, modeled using a boxcar regressor for disgust and motion-related re- with the exception of the correlation models (FWHM 8x8x8) because of gressors as confounding variables. The regressor of interest was first the low degrees of freedom in this analysis generated considering an implicit baseline condition including the ap- petizing scenes and an activation condition including the disgusting 3. Results scenes, and was then adjusted using results of a previous group analysis of an independent sample with the same experimental setup (Pujol 3.1. Subjective ratings et al., 2018). Briefly, using a ROI based approach and MarsBar tools (Brett et al., 2002), the functional MRI signal time course from 4 pre- Prior to fMRI acquisition, PWS patients reported stronger feelings of selected regions-of-interest showing the most consistent activation in hunger than control participants (mean ± SD, patients 39.5 ± 39.4, previous research (frontal cortex, insula, amygdala and fusiform gyrus controls 6.5 ± 12.9; t = 3.6 p = .002), despite the fact that ratings [Kirby and Robinson, 2017]) were averaged in each of 15 healthy were obtained after a standard meal in both groups. participants (7 males, mean ± SD age: 26.6 ± 6.2 years) to obtain a Self-report disgust ratings in PWS patients indicated that disgusting single activation time course representing activation in core (disgust) scenes during the fMRI experiment overall evoked moderate-to-high regions. Results showed that, on average across the 8 scenes, disgust subjective disgust at a level similar to that reported by control subjects stimulation activated these regions with a 3 s delay and the total mean (group mean ± SD score in patients, 68.6 ± 33.3; controls duration of core region activation was 15 s (Pujol et al., 2018).This 59.4 ± 28.3, between-group differences, t = 1.0, p = .32). dynamic information served to modify our baseline-activation regressor to adjust both activation onset and response duration to our 8 dis- 3.2. Brain response to disgust gusting events. The dynamically adjusted regressor served then to carry out the individual (first-level) analyses in our study sample. The ex- When presented with scenes of disgust (vs. appetizing food), adults perimental paradigm is shown in Supplementary Fig. 1. with Prader Willi syndrome showed bilateral activation in distributed Resulting first-level SPM contrast images for each subject were cortical areas encompassing the frontal operculum extending toward carried forward to (second) group-level non-parametric permutation- the anterior insula, the lateral frontal cortex, medial and lateral occi- based analyses using the Statistic nonParametric Mapping (SnPM13; pital areas and the fusiform gyrus (Figs. 1 and 2). At the cortical level, http://warwick.ac.uk/snpm) toolbox available as an extension of the the pattern of activation of the PWS group was qualitatively similar to SPM package. Non-parametric interrogation was selected for examina- that of the control group, albeit less extensive. Significant between- tion of second-level statistical inference given the conservative nature group differences were found in the right superior temporal/supra- and general robustness of the procedure with relatively small sample marginal gyrus, medial occipital areas and the cerebellum (Supple- sizes (Nichols and Holmes, 2002). One-sample and two-sample t-test mentary Table 1, Figs. 1 and 2). At the subcortical level, however, the designs were used to estimate significant within- and between-group presentation of disgust-related images in the PWS group did not led to activation effects. In addition, SnPM voxel-wise linear regression any significant activation of regions that were seen robustly activated in 3 L. Blanco-Hinojo et al. NeuroImage: Clinical 21 (2019) 101662 Fig. 1. Global brain response to disgusting food scenes compared with scenes of appetizing food. The analysis was adjusted to the actual brain response of core regions characterized using dynamic information from an independent experiment (see section 2.Methods). control subjects. Specifically, significantly lower activation than in the with PWS display an abnormal pattern of neural activity in response to control group was observed mostly bilaterally in a cluster encompassing disgusting food visual representations. Whilst the brain response in the amygdala, hippocampus and parahippocampal region, basal ganglia healthy subjects sequentially engaged a variety of functional networks involving the ventral striatum, thalamus, hypothalamus and periaque- involving cortical and subcortical elements (Pujol et al., 2018), the ductal gray (PAG) matter (Supplementary Table 1 and Fig. 2). Con- pattern of activation demonstrated by PWS patients was restricted to versely, patients did not demonstrate increased activation than controls the cerebral cortex. Specifically, viewing disgusting scenes in the PWS in any region. group was followed by bilateral activation in the frontal opercular-in- Similar results were observed in the comparison with the BMI- sular region, lateral frontal cortex and visual areas. However, the neural matched control group (Supplementary Figs. 2 and 3 and response was almost absent in deep structures of the limbic brain ro- Supplementary Table 2). Overall, the pattern of activation of the BMI- bustly activated in control subjects, such as the hypothalamus, amyg- matched control group was comparable to that of normal-weight con- dala/hippocampus, and PAG. trol subjects. The anterior portion of the insula and frontal operculum are con- Fig. 3 illustrates the temporal evolution of the cortical response to sidered to contain the primary gustatory cortex (Scott and Plata- disgusting food representations identified in PWS patients. Frame by Salamán, 1999) and receive somatosensory (oral cavity) and visceral frame measurements in representative regions covering the activation (gut) inputs required for taste representation (Dagher, 2012). Imaging cycle (average across the eight scenes) revealed a gradual engagement studies indicate that neural activity in these areas responds to physical of localized cortical areas reaching a peak in the insula in the fifth gustatory stimuli but also to the sight of disgusting foods (Calder et al., frame of the fMRI acquisition (the interval 8–10s after stimulus onset) 2007) and the view of facial expressions of disgust (Krolak-Salmon and persisting broadly until the end of our temporal window (i.e., for a et al., 2003; Phillips et al., 1997). In PWS patients, the insula and total of 24 s) (Supplementary Table 3). frontal operculum were activated and a significant correlation was found such that higher subjective disgust ratings were associated with 3.3. Correlations between subjective scores and brain activation increased activation in the right anterior insula. These were the most consistent findings observed in PWS patients and suggest that com- Within the PWS group, SnPM whole-brain analyses revealed a sig- monly disgusting stimuli may generate conscious sensations in the nificant positive correlation between subjective disgust ratings during distaste/disgust domain. fMRI and brain activation in two clusters centered around the anterior The lateral-ventral frontal cortex, also significantly activated in PWS insula (pseudo t = 5.4, pFWE = 0.02; cluster peak, x = 44, y = 24, patients, constitutes a main element of the frontoparietal network in- z = −16, p = .0001) and extrastriate visual cortex (pseudo t = 4.5, volved in consciously re-orienting attention to salient sensory stimuli pFWE = 0.02; peak, x = 40, y = −76, z = 28, p = .0008). Fig. 4 shows (Corbetta et al., 2008; Downar et al., 2001). Behaviorally relevant vi- the correlation map and scatter plot illustrating these associations. No sual features, as may be those related to the quality of food, elicit ac- significant association was found between subjective hunger ratings tivation in this network (Corbetta et al., 2008; Frank and Sabatinelli, and brain activation. 2012). The brain response in subjects with PWS included the frontal element of this functional network, suggesting some implicit capability of patients to process disgusting scenes as salient stimuli in the back- 4. Discussion ground of appetizing food. However, in contrast to reference control subjects, patients demonstrated poor activation in the right temporo- The results of the present study provide evidence that individuals 4 L. Blanco-Hinojo et al. NeuroImage: Clinical 21 (2019) 101662 Fig. 2. Between-group differences in the brain response to disgusting food representations. Compared with control subjects, individuals with Prader Willi syndrome showed significantly decreased activation in several subcortical structures. Right side of the figure corresponds to the right hemisphere for both coronal and axial views. parietal junction, the posterior element of the salience network particularly sensitive to disgusting stimuli (Karama et al., 2011). A (Corbetta and Shulman, 2002), which suggests that the stimuli salience recent imaging study in healthy adults showed that the processing of processing may be incomplete. inedible food activate visual-associative brain regions to a greater de- Activations in primary and secondary visual regions are part of the gree than edible food (Becker et al., 2016). Significant visual cortex normal response to emotion-evoking visual stimuli (Phan et al., 2002; activation in our study further suggests that PWS patients distinguish Stark et al., 2007). The lateral occipital and fusiform areas appear to be edible from inedible visual representations of food. However, the early Fig. 3. Graphic illustration of the temporal evolution of the response to disgusting food representations identified in Prader Willi syndrome patients. Statistic t values of the acti- vations (axis y) are plotted against time (axis x) expressed as 2-s frames starting from stimulus onset (frame 1). The bold line in axis x indicates stimulus duration (8–10 s) of the dis- gusting scenes. Data from the 8 scenes were averaged to il- lustrate the activation cycle. The activation was measured using the t-values of the region coordinate in the cluster showing peak activation across the cycle (Supplementary Table 3).The error bars represent the standard error of the mean (SEM). 5 L. Blanco-Hinojo et al. NeuroImage: Clinical 21 (2019) 101662 Fig. 4. SnPM map of the correlation between sub- jective disgust ratings and brain activation in re- sponse to the disgusting food scenes (left panel).The scatter plot (right panel) illustrates the partial cor- relation (controlling for IQ) between self-report dis- gust ratings and activation in the right anterior in- sula during scenes evoking disgust (r = 0.71, p = .001). Values on the x- and y-axis correspond to residuals. Color bar represents pseudo-t values. The sagittal view shows the right hemisphere. component of the visual activation, occurring in healthy subjects with a other structures controlling eating (Ahmed et al., 2015; Bocchetta et al., short delay of 2 s (Pujol et al., 2018), was not present in patients. The 2015). The deficient response to disgust in our PWS patients involved initial visual response to salient stimuli is thought to be modulated and both amygdala and hypothalamus, which showed no evident tissue reinforced by deep structures including the amygdala and the PAG atrophy. Other data indicate that the origin of disgust-related symptoms (Bradley et al., 2003). No deep brain activation was observed in pa- may be notably subtle. For example, pica, as the compulsive eating of tients, which is in the same direction as the lack of the early component non-nutritive substances, is a reversible phenomenon associated with of the visual activation. iron deficiency, relatively frequent in pregnant women and pre- The processing of disgusting stimuli in healthy subjects also involves adolescents (Borgna-Pignatti and Zanella, 2016). robust activation of a number of interconnected primitive/limbic brain Noteworthy, the absence of limbic response in the PWS patients structures commonly implicated in the regulation of instinctive beha- occurred even when self-report ratings of disgust did not differ between vior (Pujol et al., 2018). They include the PAG (a pivotal element in the groups. This observed dissociation between brain response and sub- mediation of automatic responses to potential danger), the amygdala/ jective ratings in PWS is consistent with family reports and data from hippocampus (ingredients common to all basic emotions that are re- behavioral studies in PWS highlighting a bias in their eating behavior sponsible for threat context learning) and the hypothalamus (the main such that despite showing an adequate general knowledge about food center for regulating food intake) (Pujol et al., 2018). Activation was (i.e., what is and isn't considered acceptable to eat), they do not care to virtually absent in all of these structures in PWS patients in our study. eat oddly combined and visibly contaminated food items (Dykens, Therefore, it appears that PWS patients are to some extent capable of 2000). It is possible then, that they know that they mustn't eat certain processing disgusting stimuli at the cortical, conscious level, but their food items and adhere to socially acceptable responses (e.g., spoiled limbic brain systems are notably insensitive to this sort of stimulation. food is disgusting), in spite of their indiscriminate eating behavior at The lack of limbic brain response may be relevant to the extent it the instinctive level. could explain altered eating behavior and subsequent obesity in PWS, The IQ was used as a confounder in the correlation analyses as there which is considered to be a failure of satiety (McAllister et al., 2011; was a trend for patients with lower IQ to report higher disgust ratings. Tauber et al., 2014) potentially related to hypothalamic dysfunction Adding this variable, however, did not change the pattern of results. (Goldstone, 2006). The hypothalamus and the amygdala are strongly Other cognitive and behavioral measures used in our previous study on connected to each other and both to the basal ganglia/orbitofrontal obsessive-compulsive phenomena in PWS (Pujol et al., 2016) particu- loops that process satiety signals. We have recently reported abnormal larly related to the inhibitory control of behavior, such as the presence functional connectivity between basal ganglia/orbitofrontal loops and and severity of compulsions, self-picking and hyperphagic symptoms, both the hypothalamus and the amygdala associated with obsessive were also considered here in relation to disgust, but we did not find any eating behavior in PWS patients (Pujol et al., 2016). Although the ex- significant association. By contrast, in the previous study we did ob- ploration of anatomical data was outside the scope of this study, it is serve a correlation between inhibitory control alterations in patients worth mentioning here that volumetric results from an ongoing in- (e.g. skin picking, obsessive eating behavior) and abnormal functional vestigation by our group examining praxis skills in the same subjects (in connectivity in pathways connecting frontal to subcortical structures preparation), show a pattern of extensive gray matter reduction in PWS involved in different aspects of executive control. Future imaging stu- patients involving predominantly the cerebral cortex. However, the dies may help to address whether other clinical expressions of the as- subcortical structures showing low response to disgusting food were sorted PWS are associated with particular brain activation patterns. notably preserved with little overlap between functional and anatomic The participants' internal state, particularly regarding hunger levels, alterations (Supplementary Fig. 4), which emphasizes the relative may also influence the motivational value of the food representations specificity of the current findings. (i.e., automatic tendencies of withdrawal) and thus modulate the feel- We would succinctly mention the commonality of our findings with ings of disgust. As revealed by previous studies investigating the spe- other neurodegenerative disorders. At the extreme of disgust-related cific influence of homeostatic dysregulation on disgust, hungry subjects disorders, coprophagia is associated with a range of neuropsychiatric exhibit weaker disgust reaction toward unpalatable foods than satiated disorders, especially in demented patients with evidence of medial participants (Hoefling et al., 2009). In our study, all participants re- temporal lobe and amygdala atrophy (Josephs et al., 2016). Damage at ceived a standardized meal an hour prior to fMRI. Although hunger both locations is in fact the neuroanatomical substrate of the Klüver- ratings immediately before the scanning were significantly higher in Bucy syndrome that is characterized by hyperorality in association with PWS patients, group mean scores were low in both cases and no sig- other compulsive symptoms and generally diminished fear responses nificant association was found between the evoked brain activation in (Lanska, 2018). However, in the behavioral variant of frontotemporal PWS and self-report ratings of hunger. dementia, abnormal eating behavior has also been related to hypotha- Some limitations of this study should be considered. Because our lamic degeneration, with potential involvement of its connections with visual stimulus featured food-related dynamic scenes continuously 6 L. Blanco-Hinojo et al. NeuroImage: Clinical 21 (2019) 101662 throughout the experiment, we cannot infer from our results that sub- frontotemporal dementia. J. Neurol. 262, 2635–2642. cortical areas showing lack of significant activation in PWS (e.g., Borgna-Pignatti, C., Zanella, S., 2016. Pica as a manifestation of iron deficiency. Expert. Rev. Hematol. 9, 1075–1080. amygdala, hypothalamus) are always inactive, but only that they are Bradley, M.M., Sabatinelli, D., Lang, P.J., Fitzsimmons, J.R., King, W., Desai, P., 2003. not more active during disgusting than during appetizing food scenes. Activation of the visual cortex in motivated attention. Behav. Neurosci. 117, Based on results of previous studies in PWS showing a strong emotional 369–380. Brett, M.A., Anton, J.L., Valabregue, R., Poline, J.B., 2002. Region of interest analysis response to food images relative to control subjects (Hinton et al., 2006; using an SPM toolbox. NeuroImage 16, 497 (abstract). Miller et al., 2007; Key and Dykens, 2008), it is likely that the limbic Calder, A.J., Beaver, J.D., Davis, M.H., van Ditzhuijzen, J., Keane, J., Lawrence, A.D., system is insensitive to the disgusting attributes of food, but no to food 2007. Disgust sensitivity predicts the insula and pallidal response to pictures of dis- gusting foods. Eur. J. Neurosci. 25, 3422–3428. representations themselves. Future studies using an additional baseline Cassidy, S.B., Schwartz, S., Miller, J.L., Driscoll, D.J., 2012. Prader-Willi syndrome. condition (e.g. neutral stimuli) may help elucidating this question. In Genet. Med. 14, 10–26. addition, a limitation when assessing patients with a degree of in- Chapman, H.A., Anderson, A.K., 2012. Understanding disgust. Ann. N. Y. Acad. Sci. 1251, 62–76. tellectual disability using fMRI is the risk of excessive head motion that Corbetta, M., Shulman, G.L., 2002. Control of goal-directed and stimulus-driven attention may reduce activation detection sensitivity. We have stressed on this in the brain. Nat. Rev. Neurosci. 3, 210–215. point adopting several procedures in order to control for the effect of Corbetta, M., Patel, G., Shulman, G.L., 2008. The reorienting system of the human brain: motion, including a rigorous exclusion of individuals with less optimal from environment to theory of mind. Neuron 58, 306–324. Dagher, A., 2012. Functional brain imaging of appetite. Trends Endocrinol. Metab. 23, image quality. 250–260. Dale, A.M., Buckner, R.L., 1997. Selective averaging of rapidly presented individual trials 4.1. Conclusion using fMRI. Hum. Brain Mapp. 5, 329–340. Dimitropoulos, A., Schultz, R.T., 2008. Food-related neural circuitry in Prader-Willi syndrome: response to high- versus low-calorie foods. J. Autism Dev. Disord. 38, Our study provides novel insights into the potential origin of eating 1642–1653. behavior alterations and subsequent obesity in patients with PWS by Downar, J., Crawley, A.P., Mikulis, D.J., Davis, K.D., 2001. The effect of task relevance on the cortical response to changes in visual and auditory stimuli: and event-related characterizing the brain response to disgusting food representations fMRI study. NeuroImage 14, 1256–1267. using fMRI. The presence of significant cortical changes further in- Dykens, E.M., 2000. Contaminated and unusual food combinations: what do people with dicates that PWS patients to some extent consciously process disgusting Prader-Willi syndrome choose? Ment. Retard. 38, 163–171. Dykens, E.M., Maxwell, M.A., Pantino, E., Kossler, R., Roof, E., 2007. Assessment of hy- stimuli. This is important to the extent that it suggests that perception perphagia in Prader-Willi syndrome. Obesity (Silver Spring) 15, 1816–1826. and the understanding of disgust are not the major handicaps in PWS Frank, D.W., Sabatinelli, D., 2012. Stimulus-driven reorienting in the ventral frontopar- patients. Results from the behavior testing were clearly in that direc- ietal attention network: the role of emotional content. Front. Hum. Neurosci. 6, 116. Goldstone, A.P., 2004. Prader-Willi syndrome: advances in genetics, pathophysiology and tion, showing that PWS patients identified disgusting food to the level treatment. Trends Endocrinol. Metab. 15, 12–20. of control subjects. Their eating control problems seem to be originated Goldstone, A.P., 2006. The hypothalamus, hormones, and hunger: alteration in human at a more basic/instinctive level of appetite control, as the virtual ab- obesity and illness. Prog. Brain Res. 153, 57–73. sence of response in the hypothalamus and related structures suggests. Hinton, E.C., Holland, A.J., Gellalty, M.S., Soni, S., Patterson, M., Ghatei, M.A., Owen, A.M., 2006. Neural representations of hunger and satiety in Prader-Willi syndrome. These results may support the potential feasibility of therapeutic in- Int. J. Obes. 30, 313–321. terventions for PWS aimed at promoting learned aversive responses Hoefling, A., Likowski, K.U., Deutsch, R., Häfner, M., Seibt, B., Mühlberger, A., Weyers, (e.g., teaching appropriate aversive reactions to the idea of eating P., Strack, F., 2009. When hunger finds no fault with moldy corn: food deprivation reduces food-related disgust. Emotion 9, 50–58. spoiled food from the garbage) in the absence of a strong negative or Holland, A.J., Treasure, J., Coskeran, P., Dallow, J., 1995. Characteristics of the eating disgust instinctive reaction. disorder in Prader-Willi syndrome: implications for treatment. J. Intellect. Disabil. Res. 39, 373–381. Holsen, L.M., Zarcone, J.R., Brooks, W.M., Butler, M.G., Thompson, T.I., Ahluwalia, J.S., Acknowledgments Nollen, N.L., Savage, C.R., 2006. Neural mechanisms underlying hyperphagia in Prader-Willi syndrome. Obesity (Silver Spring) 14, 1028–1037. This study was supported in part by the Fondo de Investigación Houben, K., Havermans, R.C., 2012. A delicious fly in the soup. The relationship between disgust, obesity, and restraint. Appetite 58, 827–830. Sanitaria del Instituto Carlos III-Fondo Europeo de Desarrollo Regional Josephs, K.A., Whitwell, J.L., Parisi, J.E., Lapid, M.I., 2016. Coprophagia in neurologic (Grant PI-14/2057) and the Fundació Parc Taulí (Grant CIR 2014/002). disorders. J. Neurol. 263, 1008–1014. The Agency of University and Research Funding Management of the Karama, S., Armony, J., Beauregard, M., 2011. Film excerpts shown to specifically elicit various affects lead to overlapping activation foci in a large set of symmetrical brain Catalonia Government participated in the context of Research Group regions in males. PLoS One 6, e22343. SGR2014-1673. We thank the Associació Catalana de Síndrome de Key, A.P., Dykens, E.M., 2008. Hungry eyes’: visual processing of food images in adults Prader-Willi, Fundación Síndrome de Prader-Willi and all the patients and with Prader-Willi syndrome. J. Intellect. Disabil. Res. 52, 536–546. their families for kindly agreeing to participate. Kirby, L.A., Robinson, J.L., 2017. Affective mapping: an activation likelihood estimation (ALE) meta-analysis. Brain Cogn. 118, 137–148. Krolak-Salmon, P., Hénaff, M.A., Isnard, J., Tallon-Baudry, C., Guénot, M., Vighetto, A., Declarations of interest Bertrand, O., Mauguière, F., 2003. An attention modulated response to disgust in human ventral anterior insula. Ann. Neurol. 53, 446–453. Lanska, D.J., 2018. The Klüver-Bucy syndrome. Front. Neurol. Neurosci. 41, 77–89. None. López-Solà, M., Pujol, J., Hernández-Ribas, R., Harrison, B.J., Ortiz, H., Soriano-Mas, C., Deus, J., Menchón, J.M., Vallejo, J., Cardoner, N., 2010. Dynamic assessment of the Appendix A. Supplementary data right lateral frontal cortex response to painful stimulation. NeuroImage 50, 1177–1187. McAllister, C.J., Whittington, J.E., Holland, A.J., 2011. Development of the eating be- Supplementary data to this article can be found online at https:// haviour in Prader-Willi Syndrome: advances in our understanding. Int. J. Obes. 35, doi.org/10.1016/j.nicl.2019.101662. 188–197. Miller, J.L., James, G.A., Goldstone, A.P., Couch, J.A., He, G., Driscoll, D.J., Liu, Y., 2007. Enhanced activation of reward mediating prefrontal regions in response to food sti- References muli in Prader-Willi syndrome. J. Neurol. Neurosurg. Psychiatry 78, 615–619. Miller, J.L., Lynn, C.H., Driscoll, D.C., Goldstone, A.P., Gold, J.A., Kimonis, V., Dykens, E., Butler, M.G., Shuster, J.J., Driscoll, D.J., 2011. Nutritional phases in Prader-Willi Ahmed, R.M., Latheef, S., Bartley, L., Irish, M., Halliday, G.M., Kiernan, M.C., Hodges, syndrome. Am. J. Med. Genet. A 155A, 1040–1049. J.R., Piguet, O., 2015. Eating behavior in frontotemporal dementia: Peripheral hor- Nichols, T.E., Holmes, A.P., 2002. Nonparametric permutation tests for functional neu- mones vs hypothalamic pathology. Neurology 85, 1310–1317. roimaging: a primer with examples. Hum. Brain Mapp. 15, 1–25. Becker, C.A., Flaisch, T., Renner, B., Schupp, H.T., 2016. Neural correlates of the per- Phan, K.L., Wager, T., Taylor, S.F., Liberzon, I., 2002. Functional neuroanatomy of ception of spoiled food stimuli. Front. Hum. Neurosci. 10, 302. emotion: a meta-analysis of emotion activation studies in PET and fMRI. NeuroImage Bocchetta, M., Gordon, E., Manning, E., Barnes, J., Cash, D.M., Espak, M., Thomas, D.L., 16, 331–348. Modat, M., Rossor, M.N., Warren, J.D., Ourselin, S., Frisoni, G.B., Rohrer, J.D., 2015. Phillips, M.L., Young, A.W., Senior, C., Brammer, M., Andrew, C., Calder, A.J., Bullmore, Detailed volumetric analysis of the hypothalamus in behavioral variant E.T., Perrett, D.I., Rowland, D., Williams, S.C., Gray, J.A., David, A.S., 1997. A 7 L. Blanco-Hinojo et al. NeuroImage: Clinical 21 (2019) 101662 specific neural substrate for perceiving facial expressions of disgust. Nature 389, 489–511. 495–498. Shapira, N.A., Lessig, M.C., He, A.G., James, G.A., Driscoll, D.J., Liu, Y., 2005. Satiety Power, J.D., Mitra, A., Laumann, T.O., Snyder, A.Z., Schlaggar, B.L., Petersen, S.E., 2014. dysfunction in Prader-Willi syndrome demonstrated by fMRI. J. Neurol. Neurosurg. Methods to detect, characterize, and remove motion artifact in resting state fMRI. Psychiatry 76, 260–262. NeuroImage 84, 320–341. Stark, R., Zimmermann, M., Kagerer, S., Schienle, A., Walter, B., Weygandt, M., Vaitl, D., Pujol, J., Macià, D., Blanco-Hinojo, L., Martínez-Vilavella, G., Sunyer, J., de la Torre, R., 2007. Hemodynamic brain correlates of disgust and fear ratings. NeuroImage 37, Caixàs, A., Martín-Santos, R., Deus, J., Harrison, B.J., 2014. Does motion-related 663–673. brain functional connectivity reflect both artifacts and genuine neural activity? Swaab, D.F., 1997. Prader-Willi syndrome and the hypothalamus. Acta Paediatr. Suppl. NeuroImage 101, 87–95. 423, 50–54. Pujol, J., Blanco-Hinojo, L., Esteba-Castillo, S., Caixàs, A., Harrison, B.J., Bueno, M., Deus, Swaab, D.F., Purba, J.S., Hofman, M.A., 1995. Alterations in the hypothalamic para- J., Rigla, M., Macià, D., Llorente-Onaindia, J., Novell-Alsina, R., 2016. Anomalous ventricular nucleus and its oxytocin neurons (putative satiety cells) in Prader-Willi basal ganglia connectivity and obsessive-compulsive behaviour in patients with syndrome: a study of five cases. J. Clin. Endocrinol. Metab. 80, 573–579. Prader Willi syndrome. J. Psychiatry Neurosci. 41, 261–271. Tauber, M., Diene, G., Mimoun, E., Çabal-Berthoumieu, S., Mantoulan, C., Molinas, C., Pujol, J., Blanco-Hinojo, L., Coronas, R., Esteba-Castillo, S., Rigla, M., Martínez-Vilavella, Muscatelli, F., Salles, J.P., 2014. Prader-Willi syndrome as a model of human hy- G., Deus, J., Novell, R., Caixàs, A., 2018. Mapping the sequence of brain events in perphagia. Front. Horm. Res. 42, 93–106. response to disgusting food. Hum. Brain Mapp. 39, 369–380. Zhang, Y., Zhao, H., Qiu, S., Tian, J., Wen, X., Miller, J.L., von Deneen, K.M., Zhou, Z., Russell, H., Oliver, C., 2003. The assessment of food-related problems in individuals with Gold, M.S., Liu, Y., 2013. Altered functional brain networks in Prader-Willi syn- Prader-Willi syndrome. Br. J. Clin. Psychol. 42, 379–392. drome. NMR Biomed. 26, 622–629. Scott, T.R., Plata-Salamán, C.R., 1999. Taste in the monkey cortex. Physiol. Behav. 67, 8