nutrients Article Sex Differences and Commonalities in the Impact of a Palatable Meal on Thalamic and Insular Connectivity Lisa Kilpatrick 1,† , Teodora Pribic 2,† , Barbara Ciccantelli 2 , Carolina Malagelada 2 , Dan M. Livovsky 3 , Anna Accarino 2 , Deborah Pareto 4 , Fernando Azpiroz 2, * and Emeran A. Mayer 1 1 Division of Digestive Diseases, G Oppenheimer Center for Neurobiology of Stress and Resilience, Los Angeles, CA 90095, USA; lisa@lisakilpatrick.com (L.K.); emayer@ucla.edu (E.A.M.) 2 Digestive System Research Unit, University Hospital Vall d’Hebron, Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (Ciberehd), Departament de Medicina, Universitat Autònoma de Barcelona, 08193 Bellaterra (Cerdanyola del Vallès), Spain; teodora.pribic@gmail.com (T.P.); bciccantelli@gmail.com (B.C.); cmalagelada@gmail.com (C.M.); aaccarino@telefonica.net (A.A.) 3 Digestive Diseases Institute, Shaare Zedek Medical Center, Hebrew University, 9103102 Jerusalem, Israel; danlivo@yahoo.com 4 Radiology Department, University Hospital Vall d’Hebron, 08035 Barcelona, Spain; dpareto@idi-cat.org * Correspondence: azpiroz.fernando@gmail.com; Tel.: +34-932-746-222 † L.K. and T.P. contributed equally as co-first authors.  Received: 28 April 2020; Accepted: 26 May 2020; Published: 1 June 2020  Abstract: The neural mechanisms underlying subjective responses to meal ingestion remain incompletely understood. We previously showed in healthy men an increase in thalamocortical, and a decrease in insular-cortical connectivity in response to a palatable meal. As sex is increasingly recognized as an important biological variable, we aimed to evaluate sex differences and commonalities in the impact of a well-liked meal on thalamic and anterior insular connectivity in healthy individuals. Participants (20 women and 20 age-matched men) underwent resting-state magnetic resonance imaging (rsMRI) before and after ingesting a palatable meal. In general, the insula showed extensive postprandial reductions in connectivity with sensorimotor and prefrontal cortices, while the thalamus showed increases in connectivity with insular, frontal, and occipital cortices, in both women and men. However, reductions in insular connectivity were more prominent in men, and were related to changes in meal-related sensations (satiety and digestive well-being) in men only. In contrast, increases in thalamic connectivity were more prominent in women, and were related to changes in satiety and digestive well-being in women only. These results suggest that brain imaging may provide objective and sex-specific biomarkers of the subjective feelings associated with meal ingestion. Keywords: meal ingestion; postprandial sensations; hedonic response; brain imaging; sex differences; resting state 1. Introduction Consumption of a meal induces a sequence of transient sensations before and during ingestion (e.g., anticipatory reward, gustation) as well as longer-lasting postprandial sensations. The postprandial experience involves homeostatic sensations (e.g., satiety) and hedonic sensations (e.g., digestive well-being), which are independently modulated by various conditioning factors [1]; for instance, more likable meals induce lower homeostatic sensations but stronger hedonic sensations [2]. The neural mechanisms underlying the postprandial experience remain incompletely understood [3]. While integration of neural and endocrine satiety signals occurs in the hypothalamus, the anterior insula plays a key role in the conscious sensation of palatability, hunger and satiation [4]. Indeed, alterations Nutrients 2020, 12, 1627; doi:10.3390/nu12061627 www.mdpi.com/journal/nutrients Nutrients 2020, 12, 1627 2 of 12 in these brain mechanisms have been reported in obesity [5]. We previously reported the results of a resting state functional magnetic resonance imaging (rsMRI) study examining the impact of a palatable meal on thalamic and insular brain networks in healthy men [6]. We found that thalamo-cortical connectivity increased and insular-cortical connectivity decreased with ingestion. Furthermore, in that study, meal-induced decreases in insular-anterior cingulate cortical (ACC) connectivity were related to more pronounced responses in self-reported satiety and digestive well-being. Sex-related differences in the biological responses to food ingestion have been reported. Particularly, women experienced stronger postprandial sensations after a palatable meal, and were more susceptible to the influence of conditioning factors, such as the eating schedule [7–9]. Sex is increasingly recognized as an important biological variable influencing the generalizability of biomedical research [10]. In particular, sex differences have been observed in the structure and function of the insula, including differences in volume and tissue density [11], in the response to an autonomic challenge [12], pain-related alterations [13], structural connectivity [14], and functional connectivity [15]. Additionally, sex differences in gray matter volume and density have been shown in the thalamus [16]. Moreover, sex differences in brain responses to food-related cues and gustatory stimuli have been reported [16–18]. Based on these data, we hypothesized that the neural mechanisms underlying the postprandial experience exhibit sex differences. Thus, our aim is to examine sex differences and commonalities in the effects of a palatable meal on brain activity, specifically, on thalamic and anterior insular connectivity, and the relations of brain responses with homeostatic sensations (satiety) and hedonic sensations (digestive well-being). 2. Methods 2.1. Participants Two groups of healthy, non-obese women and men (n = 20 each) without history of gastrointestinal symptoms participated in the study. The male participants represent a subset of the participants in a previously reported male-specific [6] and were chosen to match the women in terms of age. Since handedness influences brain organization, all participants were right-handed (as assessed by the Edinburgh Handedness Inventory) to include a homogeneous population for brain imaging studies [19]. Absence of current digestive symptoms was verified using a standard abdominal symptom questionnaire (no symptom ≥2 on a 0–10 scale); the reliability of this questionnaire has been validated by previous studies showing that it allows good discrimination between patients with functional digestive disorders and healthy subjects [20–22]. Individuals with psychological and eating disorders were excluded (as assessed by the following questionnaires: the Hospital Anxiety and Depression scale, Dutch Eating Behaviour Questionnaire, and the Physical Anhedonia Scale). Exclusion criteria were: history of anosmia, ageusia, current or recent dieting or any pattern of selective eating, such as vegetarianism, alcohol abuse, and the use of recreational drugs. For this pilot study no formal sample size calculations were performed. Participants were recruited by public advertisement in the university campus and received a monetary compensation for their participation. The study protocol, including recruitment and compensation procedures, was approved by the Institutional Review Board of the University Hospital Vall d’Hebron and was registered with Clinical Trials.gov as part of the study NCT02592239. All participants gave written informed consent. Nutrients 2020, 12, 1627 3 of 12 2.2. Study Design and Procedure To address our hypothesis we evaluated thalamic and insular connectivity by brain imaging and homeostatic and hedonic sensations by scales (outcome measures) in response to a palatable probe meal (intervention). Studies were conducted in the afternoon after a 5-h fast. Participants were instructed to refrain from strenuous physical activity on the day before the study and to have their usual breakfast after an overnight fast. Participants completed a structural MRI scan and two 20-min resting state MRI (rsMRI) scans (pre-1 and pre-2) immediately before ingestion of a probe meal, and a third 20-min rsMRI scan after ingestion (post meal); the two pre-meal rsMRI scans were acquired to examine the effect of time. Perception of subjective sensations was measured after the second fasting rsMRI scan (pre-meal), and immediately before and after the postprandial scan (0 min and 20 min post-meal, respectively). The probe meal was ingested while participants sat in an isolated room adjacent to the brain imaging room. 2.3. Intervention: Probe Meal To evaluate the brain activity underlying a positive postprandial experience, a probe meal (comfort-type meal) was specifically designed by a series of previous studies [23]; the characteristics of the meal (amount, composition, and palatability) were tailored to induce a consistent homeostatic response (satiation) with a pleasurable hedonic dimension (digestive well-being). The probe meal (425 kcal; 47 g carbohydrates, 17 g lipids, 18 g proteins) consisted of 200 mL orange juice and a warm sandwich (58 g white bread with 12 g butter, 38 g ham and 38 g cheese) freshly cooked on a hot plate (Sandwich Maxi 20, Fagor, Olite, Spain) for 3 min and administered at a standard temperature (allowed to cool down covered by a napkin for 3 min). 2.4. Outcome Measures 2.4.1. Assessment of Subjective Responses Ten centimeter scales graded from −5 to +5 were used to measure hunger/satiety (extremely hungry to completely sated), digestive well-being (extremely unpleasant sensation/dissatisfaction to extremely pleasant sensation/satisfaction), and palatability (very bad/disagreeable to very good/delicious). Subjects received standard instructions on how to complete the scales: The hunger/satiety and digestive well-being scales were scored immediately before, at the end and 20 min after meal ingestion (score the intensity of each sensation you feel now); the palatability scale (how did you find the meal) was scored only once immediately after ingestion. These scales have been shown to detect differences in postprandial sensations under various conditioning factors [1,2,9,24]. 2.4.2. Neuroimaging fMRI acquisition. Images were acquired on a 3.0 T whole-body MR scanner with a 12-channel phased-array head coil and a whole-body transmit coil (Trio, Siemens, Germany). The imaging protocol parameters were based on the ADNI initiative.The structural scan involved an axial 3D T1-weighted gradient-echo (MPRAGE) sequence (TR = 2200 ms, TE = 3.26 ms, voxel size = 1.0 × 1.0 × 1.0 mm3 ). Three 20-min rsMRI scans (pre-1, pre-2 and post ingestion) were completed, in which participants rested with eyes closed (TR = 2000 ms, TE = 28 ms, voxel size = 3.4 × 3.4 × 4.0 mm3 ). fMRI preprocessing. Imaging data were pre-processed using SPM12 software (Welcome Department of Cognitive Neurology, London, UK). The first two volumes were discarded to allow for stabilization of the magnetic field. Data were slice-time and motion corrected, spatially normalized to the Montreal Neurological Institute standard template using their structural image, and smoothed with an 8-mm Gaussian kernel. In addition, white matter/cerebrospinal fluid signals were regressed out of the time series data, which were then bandpass filtered (0.01–0.10 Hz). Nutrients 2020, 12, 1627 4 of 12 2.5. Statistical Analysis 2.5.1. Perception Measurements Mean values (±SE) of the parameters measured were calculated in each group of subjects. Normality of data distribution was evaluated by the Kolmogorov-Smirnov test. Comparisons of parametric normally distributed data were made by Student’s t test, paired tests for intragroup comparisons and unpaired tests for intergroup comparisons; otherwise, the Wicoxon signed rank test was used for paired data within groups, and the Mann-Whitney U test for unpaired data between groups. Differences were considered significant at a p value < 0.05. 2.5.2. Brain Connectivity Seed-based correlation has been used in the analysis of rsMRI data, as previously described in detail [25]. In brief, this method provides direct and easily interpreted information regarding connectivity for pre-defined regions of interest. The left and right short insula gyri and left and right thalamus, based on the Destrieux atlas [26], were utilized as regions of interest (ROIs) in the present study. After regressing out the motion parameters and bandpass filtering (0.01–0.08 Hz) to reduce noise, the mean time course was extracted from each of the ROIs, and correlated with the time courses of all other grey matter voxels. The resulting correlation maps were normalized using Fisher’s r-to-z transformation [27]. Partial least squares (PLS) analysis was applied to the normalized correlation maps to examine sex-related and meal-ingestion effects on anterior insula and thalamic connectivity, using freely available code [28]. PLS is a multivariate statistical technique that is used to find the relationship between two blocks of variables, with numerous advantages; for example, unlike SPM, PLS does not require a prior specification of the contrasts of interest, and PLS is more sensitive than traditional univariate SPM-based analyses [28,29]. Four seed-PLS analyses were conducted, one for each of the four ROIs (left and right anterior insula; left and right thalamus). In each analysis, sex was a group factor and there were three meal-ingestion conditions (pre-1, pre-2, and post-meal); the normalized correlation maps were provided as input. As a result, each seed-PLS analysis produced a set of spatial maps with associated weightings for each combination of sex and condition (i.e., latent variables). Significance of each latent variable was assessed using permutation testing (500 permutations), and voxel reliability was assessed using bootstrap estimation (500 samples) [28]. Clusters with a peak voxel bootstrap standard error ratio (BSR) exceeding ±3.3 (p < 0.001) and an extent of at least 150 voxels were considered as reliable. 2.5.3. Correlation of Brain Connectivity and Sensations Reliable clusters from the seed-PLS analyses were further analyzed in focused behavioral-PLS analyses, which evaluated the relationships between connectivity changes and changes in perceptual ratings within each sex. As the seed-PLS analyses showed no significant differences between the first and second pre-meal ingestion scan, connectivity difference maps were calculated as the post-meal ingestion connectivity map minus the second pre-meal ingestion connectivity map (post minus pre-2). The ingestion-related changes in satiety and digestive well-being ratings (∆ satiety and ∆ well-being) were calculated as the post-meal ingestion rating minus the pre-meal ingestion rating. For each sex, four behavioral-PLS analyses were conducted, one for each of the four ROIs; the connectivity difference maps and both ∆ satiety and ∆ well-being were provided as input. As a result, each behavioral-PLS analysis produced a set of spatial maps with associated weightings for ∆ satiety and ∆ well-being (i.e., latent variables). Significance of each latent variable and voxel reliability were assessed in the same manner as in the seed-PLS analyses. Nutrients 2020, 12, 1627 5 of 12 3. Results 3.1. Demographics Age range was similar in women (20–43 years) and men (20–40 years). Body mass index was also similar inx women 2 2 Nutrients 2017, 9, FOR PEER(range REVIEW19–29 5 of 12Kg/m ) and men (20–29 Kg/m ); one participant in each group was overweight and the rest were within the normal range. Participants were university students or had aa university had university degree degree (middle-upper (middle-upperclass) class)and andall allcompleted completedthe thestudy studyand andwere wereincluded includedinin the the final analysis. final analysis. 3.2. Meal-Related Meal-Related Sensations Sensations Both men and women women found found the the probe probe meal meal equally equally likeable likeable(palatability (palatabilityscore score3.6 3.6± ±0.20.2inin women and 3.9 ± 0.2 in men; p = women = 0.210). 0.210). Meal Mealingestion ingestionincreased increasedsatiety satietyand anddigestive digestivewell-being well-being in men and women (p < 0.001 for all) and these effects persisted over the 20-min postprandial ratings in men and women (p < 0.001 for all) and these effects persisted over the 20-min postprandial period (p period (p<< 0.001 0.001for forall) all)(Figure (Figure1).1).InInwomen, women,satiety satiety ratings ratings at at 0 and 0 and 20 min 20 min after after meal meal ingestion ingestion were were higher higher than inthan = 0.003 menin(pmen (p =for 0.003 both)for both) 1A). (Figure (Figure Sex 1differences A). Sex differences were alsowere also observed observed in relationinto relation to digestive digestive before well-being: well-being: before the meal, the meal, well-being well-being ratings ratings were higher werethan in men higher in men(pthan in women in = 0.048), women (p = 0.048), but ratings immediately after the meal were higher in women than but ratings immediately after the meal were higher in women than in men (p = 0.006); hence, the increase in men (p = 0.006); in hence, (∆ well-being thewell-being) increase in was well-being larger in(Δwomen well-being) thanwas larger in men in women (Figure 1B). Atthan in men 20 min after(Figure ingestion 1 B). At 20 min after ingestion well-being rates were well-being rates were similar in women and men (Figure 1B) similar in women and men (Figure 1 B) Figure 1. Sensory responses to meal ingestion. Satiety (A) and digestive well-being (B) ratings (mean values Figure and SE)1.before Sensory responses (Pre) and at 0toand meal20ingestion. min afterSatiety meal (A) and digestive ingestion. well-being Note more (B) ratings prominent (mean homeostatic values and SE) before (Pre) and at 0 and 20 min after meal ingestion. Note more response (satiety) and early hedonic response (digestive well-being) in women than in men.* p < 0.05prominent homeostatic vs. response Pre (significant (satiety) and early ingestion-related hedonic change); † presponse (digestive < 0.01 vs. well-being)sex men (significant in difference women than in within men.* p condition).< 0.05 vs. Pre (significant ingestion-related change); † p < 0.01 vs. men (significant sex difference within condition) . 3.3. Brain Imaging 3.3. Brain Imaging No significant differences between the two pre-meal ingestion scans were detected both in women No significant and men, differences indicating the between lack of time effectsthe twothe before pre-meal ingestion intervention (mealscans were in ingestion); detected contrast,both in marked women and and men, consistent indicating changes the lackafter were detected of time mealeffects before the intervention (meal ingestion); in ingestion. contrast, marked and consistent changes were detected after meal ingestion. 3.3.1. Anterior Insular Connectivity Significant connectivity patterns (latent variables) were detected for right anterior insular (accounting for 72.3% of the cross-block variance; p < 0.001) and for left anterior insula (accounting for 66.1% of the cross-block variance; p < 0.001). Before the meal, i.e., at baseline, significant sex differences in connectivity for the right and left anterior insula were observed (Figure 2A and B): specifically, connectivity between the anterior insula and sensorimotor and prefrontal cortices was greater in men than in women during both pre-1 scan and pre-2 scan (Table 1 and Figure S1-A and S1-B). In men meal ingestion significantly reduced connectivity of the left and right anterior insula with these regions; in women the change was not significant and tended to be smaller (although not Nutrients 2020, 12, 1627 6 of 12 3.3.1. Anterior Insular Connectivity Significant connectivity patterns (latent variables) were detected for right anterior insular (accounting for 72.3% of the cross-block variance; p < 0.001) and for left anterior insula (accounting for 66.1% of the cross-block variance; p < 0.001). Before the meal, i.e., at baseline, significant sex differences in connectivity for the right and left anterior insula were observed (Figure 2A,B): specifically, connectivity between the anterior insula and sensorimotor and prefrontal cortices was greater in men than in women during both pre-1 scan and pre-2 scan (Table 1 and Figure S1A,B). In men meal ingestion significantly reduced connectivity of the left and right anterior insula with these regions; in women the change was not significant and tended to be smaller (although not significantly) than in men, so that sex differences in connectivity observed at baseline were no longer detectable after meal ingestion Nutrients 2017, 9, x FOR PEER REVIEW 6 of 12 (Figure 2A,B). Figure 2. Brain responses to meal ingestion. Brain connectivity summary scores (mean and confidence Figure interval)2. Brain in theresponses to meal two sequential ingestion. scans beforeBrain connectivity the meal (pre-1 and summary scores pre-2) and (mean in the scanand confidence after the meal interval) (post) areinplotted the two sequential according to scans sex andbefore the meal condition for (pre-1 and pre-2) each region and in(A) of interest: theLeft scananterior after theinsula; meal (post) are plotted according to sex and condition for each region of interest: (A) Left anterior (B) right anterior insula; (C) left thalamus; and (D) right thalamus. The brain connectivity summary insula; (B) right scores anterior reflect insula; (C) (analogous the weightings left thalamus; and (D) to factor rightinthalamus. scores The brain a factor analysis) connectivity determined summary by the partial scores reflect the weightings (analogous to factor scores in a factor analysis) determined by least squares analysis. The brain regions reliably contributing to these weightings are shown in Figure the partial S1. * p < squares least 0.05 vs. analysis. The brain Pre (significant regions reliablychange); ingestion-related † p