Chapter 10 Discussion 2

Dr. Bedwell published a recent paper that showed that UCF undergraduate women who were in their mid-to-late luteal phase of the menstrual cycle had a reduction in reported pleasure during a behavioral task as compared to the other women and men. See .pdf link below for the paper. Please summarize what is discussed about the likely hormone that was responsible for this finding (see Discussion section) and what is known about the effect of this hormone on pleasure from other studies reviewed in the Introduction and Discussion sections of below paper.


200 word count

The Sweet Taste Test: Relationships with Anhedonia Subtypes, Personality Traits, and Menstrual Cycle Phases

Jeffrey S. Bedwell1 & Christopher C. Spencer1 & Cristina A. Chirino1 & John P. O’Donnell1

Published online: 15 January 2019 # Springer Science+Business Media, LLC, part of Springer Nature 2019

Abstract A better understanding of the etiology and pathology related to distinct subtypes of anhedonia can lead to more efficacious personalized treatments. The current study advances knowledge on consummatory anhedonia – represented in the Research Domain Criteria (RDoC) subconstruct of Initial Response to Reward (IRR). The Sweet Taste Test (STT) has promise as a behavioral paradigm for IRR, as it is sensitive to manipulation of μ-opioid receptors. However, there is a lack of existing knowledge of how the STT relates to subtypes of anhedonia, personality traits, and phases of the menstrual cycle. To address these questions, we administered the STT to 72 nonpsychiatric adults (76% women; mean age: 19.11). As predicted, the hedonic slope reflecting increasing Blike^ ratings over increasing concentrations of five sucrose solutions (ranging from 0.05 M to 0.86 M), was lower in individuals reporting higher consummatory anhedonia (measured with Temporal Experience of Pleasure Scales – Consummatory Subscale) and in women in the mid-to-late luteal menstrual phase (days 20 to 28). Both effects were driven by lower hedonic ratings to the sweetest concentration. The hedonic slope was larger in individuals scoring higher on the Flight-Freeze-Avoidance System personality factor from the Reinforcement Sensitivity Theory Personality Questionnaire – driven by lower hedonic ratings for the least sweet concentration. No factor or aspect from the Big Five Aspects Scale related to hedonic ratings on the STT. The STT may be a valid and specific standardized behavioral paradigm to add to IRR, particularly if validated in a large transdiagnostic psychiatric sample.

Keywords Sucrose .Reward .Consummatoryanhedonia .Initialresponsivenesstorewardattainment .Researchdomaincriteria .

Consummatory pleasure


Anhedonia is a transdiagnostic symptom which is particularly resistant to existing treatments relative to most co-occurring symptoms (McCabe et al. 2010; Vittengl et al. 2015). While the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) defines anhedonia as a unitary construct (e.g., Bdecreased interest and pleasure in most activities most of the day;^ American Psychiatric Association 2013), recent ad- vances in affective neuroscience suggest three distinct sub- types (Treadway and Zald 2011): motivational (i.e., wanting), consummatory (i.e., liking), and decisional (i.e., reward learn- ing). The National Institute of Mental Health’s (NIMH) Research Domain Criteria (RDoC) Matrix includes

corresponding constructs under the Positive Valence Systems (PVS) domain (Insel et al. 2010). A better under- standing of the etiology and pathology related to distinct sub- types of anhedonia can lead to more efficacious personalized interventions for this treatment-resistant symptom (Insel and Cuthbert 2015; Strauss and Cohen 2017).

Consummatory anhedonia is partially represented in the PVS construct Reward Responsiveness under the subconstruct of Initial Response to Reward (IRR; as of 10/2/ 18), defined by NIMH as: BProcesses evoked by the initial presentation of a positive reinforcer as reflected by indices of neuronal activity and verbal or behavioral responses.^ This is reflected by the inclusion of the Consummatory sub- scale of the Temporal Experience of Pleasure Scales (TEPS-C; Gard et al. 2007) under Self-Reports in IRR. While IRR cur- rently names a single paradigm of Bsimple guessing task,^ application of such tasks require physiological measures of brain response (e.g., Tsypes et al. 2018). Alternatively, the Sweet Taste Test (STT; Kampov-Polevoy et al. 1997; Dichter et al. 2010) has promise as a behavioral IRR paradigm

* Jeffrey S. Bedwell

1 Department of Psychology, University of Central Florida, 4111 Pictor Lane, Orlando, FL 32816-1390, USA

Journal of Psychopathology and Behavioral Assessment (2019) 41:235–248



that could be used in clinical settings, as it is does not require physiological equipment and related expertise. While the STT is not specifically named under Paradigms for IRR, Btaste reactivity^ is listed under Behavior. In the STT, the participant tastes a random order of five trials for each of five concentra- tions of sucrose, ranging from minimal sweetness to two so- lutions that are sweeter than beverages such as Coca-Cola®, and collects subjective ratings of sweetness and liking after each trial (Kampov-Polevoy et al. 1997).

Individuals whose hedonic Blike^ ratings are highest for the strongest sucrose concentration have been termed Bsweet likers,^ a category associated with alcohol use dis- order (Kampov-Polevoy et al. 2001, 2003; Wronski et al. 2007), cocaine use disorder (Janowsky et al. 2003), binge- eating disorder (Goodman et al. 2017), impulsive choice on a delay discounting task (Weafer et al. 2014), better response to naltrexone treatment for alcohol use disorder (Garbutt et al. 2009, 2016), and greater amphetamine- induced euphoria in nonpsychiatric women but not men (Weafer et al. 2017). Consistent with inclusion of μ- opioid receptors under IRR Molecules, research has shown that stimulation of μ-opioid Bhedonic hotspots^ in the nucleus accumbens shell, ventral pallidum, anterior orbitofrontal cortex, and posterior insula increase sucrose Bliking^ behavior in rats (Castro and Berridge 2017). Similarly, a study of nonpsychiatric men found a decrease in sucrose liking following administration of the μ-opioid antagonist naltrexone, and increased liking following the μ-opioid agonist morphine (Eikemo et al. 2016). Both effects were observed for only dimensional hedonic rat- ings for the sweetest concentration (0.65 M in that study) and did not change a slope-based categorical sweet liker status. This finding is consistent with studies showing that naltrexone reduced average STT hedonic ratings in non- psychiatric women (Arbisi et al. 1999) and opioid depen- dent participants (Langleben et al. 2012). Overall, it ap- pears that the STT partially reflects endogenous opioid functioning, particularly the dimensional hedonic rating to the sweetest solution.

Considering this emerging literature, we were interested in how STT performance relates to subtypes of anhedonia and a range of personality traits. This will inform investigators and clinicians regarding which anhedonia measure or subscale may best assess IRR/consummatory anhedonia, and general personality research by highlighting particular traits which may be sensitive to individual differences in the endogenous opioid system. Finally, there is increasing evidence that the luteal menstrual cycle phase, via an increase in progesterone, dampens hedonic responsivity to cocaine in female rhesus macaques (Carroll et al. 2016) and humans (Evans and

Foltin 2006; Evans et al. 2002). As both progesterone and estradiol interact with opioid receptors (Schroeder et al. 2003; Lee and Ho 2013), it is possible that estradiol, which is higher during the follicular menstrual phase, may also effect reward sensitivity. However, studies in rhesus monkeys found that administering exogenous estradiol did not alter cocaine self-administration, while administering progesterone reduced the cocaine use (Mello et al. 2008, 2011). A recent study found an amplitude reduction in the event-related potential of reward positivity in response to monetary gains in women during the luteal as opposed to follicular phase of the menstru- al cycle, but this influence was only significant in women with greater severity of depression (Mulligan et al. 2018). Examining the relationship between menstrual phases and STT performance in a nonpsychiatric sample will account for important individual differences present in menstruating women and add to our understanding of how these phases may induce temporary changes in IRR.

It appears that only one existing study examined relation- ships between a behavioral assessment of sweet taste liking and anhedonia severity. This study found that, across a transdiagnostic sample of individuals with major depressive disorder, schizophrenia, and nonpsychiatric controls, higher scores on the self-report Chapman Physical, but not Social, Anhedonia scale was associated with a lower maximum hedonic rating across all sucrose concentrations on the STT (Berlin et al. 1998). There appear to be few published studies on the relationship of a behavioral measure of sweet liking with personality traits. One reported that individuals who preferred the sweeter of two blinded samples of the same white wine scored lower on self-reported openness and higher on impulsiveness, a facet of neuroticism (Saliba et al. 2009). A study using five blinded levels of sweetness of the same red wine found that participants who preferred sweeter concentrations were higher in neu- roticism (Sena-Esteves et al. 2018). A study using the STT reported that categorical sweet likers did not differ from other participants on self-reported novelty seeking (Lange et al. 2010). Studies that did not include a behavioral mea- sure found that higher self-reported generalized liking of sweet foods and drinks was associated with higher self- reported agreeableness (Sagioglou and Greitemeyer 2016; Meier et al. 2012; Ashton et al. 2014), neuroticism (Kikuchi and Watanabe 2000; Elfhag and Erlanson-Albertsson 2006), and/or openness/intellect (Ashton et al. 2014) scores. Finally, while some research has found that women con- sume more sweet foods during the luteal phase (Bowen and Grunberg 1990; Tucci et al. 2010), one study examined hedonic rating to varying concentrations of sucrose solu- tions reported that ratings were lowest in the luteal phase,

236 J Psychopathol Behav Assess (2019) 41:235–248



particularly for the sweetest concentration (Elliott et al. 2015). However, this study did not use the traditional STT, was limited by a single trial of four sucrose solutions, and did not include men in the comparison group.

Based on existing findings using behavioral measures of hedonic response to sweet tastes, and current information in the IRR domain, we hypothesized that hedonic ratings to the sweetest STT concentration would be lower in participants reporting increased consummatory anhedonia, and higher in participants reporting more neuroticism and/or impulsiveness. In addition, we predicted that women in the luteal phase of the menstrual cycle would report a reduced hedonic response to the sweetest STT concentration as compared to other women and men.



Participants were undergraduate students enrolled in a Psychology Department course that offered credit in ex- change for research participation at a large southeastern public university in the United States. Participants were screened online (N = 2135) and excluded for: completing questionnaire either too quickly (< 10th %ile of duration from sample; N = 122; 5.7%) or slowly (> 90%ile; N = 146; 6.8%), scoring >2 SD above mean on the Abbreviated Marlow-Crowne Social Desirability Scale (Reynolds 1982; N = 75; 3.5%), using non-prescribed stim- ulant or narcotic medication (N = 23; 1.1%), excessive chronic alcohol use (N = 16; 1.0%), hypothyroidism in self (N = 30; 1.4%) or first-degree family member (N = 91; 4.3%), significant past head injury or chronic neurological disorder (N = 50; 2.3%), failure to endorse willingness to abstain from recreational drugs for 48 h prior to the testing session (N = 369; 17.3%) or alcohol for 24 h prior to testing (N = 11; 0.5%), significant uncorrected vision impairment (N = 51; 2.4%), physical impairment in arms/hands (N = 4; 0.2%), or endorsing more than two items incorrectly on an 8-item Infrequency Scale (Jackson 1984; N = 16; 1.0%). The remaining 1131 participants received an invitation to participate in the lab-based portion of the task. However, only a minority of those participants completed the lab ses- sion (N = 79), as the online questionnaires were designed for a larger study examining relationships between those scales (currently unpublished). On the day of laboratory testing, all participants denied recent nicotine use, alcohol in past 24 h, and other recreational drug use in past 48 h.

However, this was based on self-report as we did not have laboratory tests to confirm.

Of the 79 participants, 77 had valid data for the STT, as two were excluded for sweetness ratings that did not approximate a linear increase with increasing concentrations and were sta- tistical outliers for that slope value (Z < −3.33 for each). An additional 5 participants were excluded for missing all anhe- donia measures and bio-demographic data (e.g., menstrual phase) due to experimenter error (both a priori reasons for excluding participants). This resulted in 72 participants used in final analyses (76% women; mean age: 19.11; SD = 2.07; range 18 to 28). For race, 75.0% identified as BWhite,^ 8.3% as BBlack,^ 8.3% as BAsian, 4.2% as BOther,^ 2.8% as BMixed,^ and one participant declined to answer. Independent of race, 34.7% identified as BHispanic/ Latino(a).^ Two participants reported selective-serotonin re- uptake inhibitor (SSRI) medication use at the time of testing. The remaining participants denied current psychotropic and narcotic medication use.


Self-Report Scales

The Reinforcement Sensitivity Theory of Personality Questionnaire (RST-PQ; Corr and Cooper 2016) was admin- istered during the online phase, which consists of 65 state- ments with a response key asking the participant to rate how much the statement describes them in general – ranging from B1 – Not At All^ to B4 – Highly.^ The RST-PQ produces six scales – four behavioral activation system subscales: Reward Interest, Goal-Drive Persistence, Reward Reactivity, and Impulsivity, along with scales for the Behavioral Inhibition System (BIS) and Flight-Freeze- Avoidance System (FFAS). The reliability and validity of the RST-PQ have received ad- ditional recent support in chronic pain samples (Amiri et al. 2017), and through relationships found with event-related po- tentials from electroencephalogram (Kaye et al. 2018; De Pascalis et al. 2017). The Big Five Aspect Scales (BFAS; DeYoung et al. 2007) was also administered online, and con- sists of 100 statements with a response key asking the partic- ipant to rate how accurately the statement reflects how they generally are – ranging from B1 – Very Inaccurate^ to B5 – Very Accurate.^ The BFAS results in the traditional big five factors along with two aspects for each factor: Neuroticism (Volatility, Withdrawal), Agreeableness (Compassion, Politeness), Conscientiousness (Industriousness, Orderliness), Extraversion (Enthusiasm, Assertiveness), and Openness/Intellect (Openness, Intellect). The validity of BFAS aspect scales have been supported with specific

J Psychopathol Behav Assess (2019) 41:235–248 237



relationships to psychopathology (Allen et al. 2018; Quilty et al. 2014) and self-reported personality disorder symptoms (DeYoung et al. 2016).

During the day of STTassessment, we administered three anhedonia self-report scales – the 18-item Temporal Experience of Pleasure Scales, which includes subscales for anticipatory (TEPS-A) and consummatory (TEPS-C) anhedonia (Gard et al. 2007), the 17-item Anticipatory and Consummatory Interpersonal Pleasure Scale total score (ACIPS; Gooding and Pflum 2014), and the 14-item Snaith-Hamilton Pleasure Scale total score (SHAPS; Snaith et al. 1995). Unlike the TEPS, factor analysis has not supported separate factors for anticipatory and consum- matory anhedonia on the ACIPS (Gooding and Pflum 2014). Increased anhedonia is indicated by lower scores on the TEPS and ACIPS and higher scores on the SHAPS. The Cronbach alphas for the anhedonia scales in our sample were: TEPS-A (.76), TEPS-C (.75), ACIPS (.91), and SHAPS (.59).

Sweet Taste Test (STT)

The STT is a standardized assessment of hedonic response to sweet tastes in human participants (Kampov-Polevoy et al. 1997). Powdered sugar and distilled water were used to create five concentrations of sucrose solutions (approximately 0.05 M, 0.10 M, 0.19 M, 0.42 M, and 0.86 M). As a reference, Coca-Cola® is a 0.33 M solution. Participants received five trials of each of the five solutions in a blinded random order. During each trial, participants drank from a small plastic opaque cup containing 2 mL of solution, swished the solution in their mouth for about 5 s, and then expectorated into a container. Participants were asked to mark an BX^ on a 201 mm analog scale line to indicate perceived sweetness, from BNot sweet at all^ (left anchor) to BExtremely sweet^ (right anchor), and, on a separate scale, how much they liked it from BDislike it very much^ (left anchor) to BLike it very much^ (right anchor). Following the ratings, participants rinsed with distilled water and proceeded to the next trial. Ratings were then scored based on distance (in mm) of the BX^ from the left side of the respective scale line, and then averaged within the five concentrations.

We used the slope of hedonic ratings by increasing sucrose concentration (i.e., hedonic slope) as the primary measure of dimensional hedonic ratings and any statisti- cally significant relationship with this dependent variable was then further explored across the separate hedonic rat- ings for the five concentrations. We also examined rela- tionships with the categorical Bsweet liker^ status, defined as having the highest hedonic rating for the sweetest

concentration (Kampov-Polevoy et al. 2001, 2003; Wronski et al. 2007). Finally, we included the sweet sen- sitivity slope to clarify specificity of any findings with the hedonic variables. The hedonic and sweet sensitivity slopes and the five individual concentration hedonic rat- ing values all showed relatively normal distributions with skewness and kurtosis <1.05.

Menstrual Cycle Stage Identification

We asked all female participants to identify the current day in their menstrual cycle using day one defined by the start of last menstruation. We also asked women if they thought they may be pregnant, which all denied. We created a Bluteal menstrual phase status^ variable which classified women reporting being in the mid-to- late luteal menstrual stage during the day of testing, de- fined as days 20 to 28 from start of last mentruation, with all remaining women and men combined in the second of the two categories. Although reward processing studies with rhesus macaques typically chose a late luteal window around days 24 to 27 to optimally isolate the rise in pro- gesterone (Carroll et al. 2016), we chose a wider mid-to- late luteal phase window to allow for inherent error in estimating the current day of cycle based on self-report/ memory, and to allow for inclusion of a sufficient number of females in this subgroup for statistical analyses. Similarly, we created a Bfollicular menstrual phase status^ variable which classified women being in the mid-to-late follicular phase during the day of testing, which we de- fined as days 5 to 12 following the onset of menstruation. This window is also somewhat wider than the follicular window typically used in the rhesus monkey studies (7 to 10 days following menstruation; Carroll et al. 2016), based on the same rationale.


This study was approved by the Institutional Review Board of the authors’ university. Research followed ethical principles described in the Declaration of Helsinki. Participants provided informed consent at the beginning of both online and lab- based sessions. During the lab-based session, participants completed demographic information, including questions about current medication and supplement use, the anhedonia questionnaires, and the STT, as part of a larger study. No data resulting from the larger study have been published at this time.

238 J Psychopathol Behav Assess (2019) 41:235–248



Statistical Analyses

For self-report measures, one participant was missing the TEPS and ACIPS. Across measures, if either one or two items for a subscale/factor were skipped by a particular participant, the respective score was then based on the average of the remaining items for that factor. IBM SPSS Statistics (Version 25) was used for all analyses. Initial regressions examined relationships of the STT he- donic and sweet sensitivity slopes and sweet liker catego- ry with simultaneous entry of: age, sex, luteal menstrual phase status, follicular menstrual phase status, birth con- trol medication use, race, ethnicity, and time of day during testing (variance inflation factor [VIF] < 1.60). Time of day during testing was rounded to nearest hour (mode = 4:00 p.m.; range = 9:00 a.m. to 4:00 p.m.). This was ex- amined as an approximation of fluctuations in the circa- dian system (e.g., hormone levels), which can influence behavioral responses to rewards (DePoy et al. 2017).

Additional regressions were used to examine sets of predictors on each of the three STT variables, the: 1) four anhedonia scales: TEPS-A, TEPS-C, ACIPS, and SHAPS, 2) six factor scores from the RST-PQ, 3) five BFAS fac- tors, and 4) ten BFAS aspects. Simultaneous entry was used for all predictor sets (VIF < 3.10 for all sets). If a predictor showed a significant relationship with the he- donic slope, this was followed by an exploratory analysis of the five concentration hedonic ratings that comprised the slope, using stepwise entry, regressed on the respec- tive predictor. All regressions involving the sweet liker category used binary logistic regressions, while the re- maining analyses on dimensional STT variables used

linear regression. Across all regressions, there were no statistical outliers, as defined by a Studentized residual >3.0.


See Fig. 1 for a depiction of the hedonic slope by sweet liker status. Descriptive statistics and zero-correlations for STT var- iables with each of the predictor sets can be found in Tables 1 (anhedonia scales), 2 (RST-PQ scales), 3 (BFAS factors), and 4 (BFAS aspects). Thirty participants (42%) were classified as Bsweet likers.^ Excluding the two participants taking SSRI medications did not change the pattern of significant findings across all analyses. They were therefore retained in all results presented below.

STT Relationships with Menstrual Phase Status, Demographic Variables, and Time of Day during Testing

We examined the three STT variables using regressions with simultaneous entry of sex, luteal menstrual phase status (N = 13), follicular menstrual phase status (N = 13), birth control medication use (N = 19), age, race, ethnicity, and time of day during testing. As shown in Table 5, women in the mid-to-late luteal m