Hypothesis 1: group differences in anticipatory and consummatory pleasureFull TEPS subscales

Consistent with our hypothesis, the ARFID group endorsed significantly lower anticipatory pleasure, t(102) = − 2.30, p = 0.02, d = 0.49, and consummatory pleasure, t(102) = − 3.33, p = 0.001, d = 0.70, than the healthy control group.

However, in contrast to our predictions, group differences in pleasure did not remain significant when controlling for depression and BMI. The overall model assessing group differences in anticipatory pleasure was significant, F(3, 99) = 6.41, p < 0.001, but depression was the only variable with a significant effect on anticipatory pleasure, F(1, 99) = 13.34, η2 = 0.11, p < 0.001. In the ANCOVA assessing group differences in consummatory pleasure, the overall model was significant, F(3, 99) = 5.81, p = 0.001 and depression was the only variable with a significant effect F(1, 99) = 6.10, η2 = 0.05, p = 0.02 (see Table 2 for estimated marginal means of anticipatory and consummatory pleasure for each group). The results from these models suggest that individuals with greater depression symptoms experience decreased anticipatory and consummatory pleasure, regardless of diagnostic status and BMI.

Table 2 Estimated marginal means of anticipatory and consummatory pleasure when controlling for age, depression, and body mass indexTEPS subscales without food items

In contrast with our hypothesis, there were no group differences in anticipatory pleasure, t(102) = − 1.62, p = 0.11, and consummatory pleasure, t(102) = − 1.13, p = 0.26, when food items were omitted from the TEPS. Because the ARFID and HC groups did not differ significantly on either TEPS subscale, we did not perform follow-up analyses assessing for the effects of depression and BMI.

Hypothesis 2: relationship between ARFID severity and pleasureFull TEPS subscales

The correlation coefficients between ARFID severity, anticipatory pleasure, and consummatory pleasure are presented in Table 3. ARFID severity (as measured by the PARDI severity scale) was positively correlated with depression, r = 0.43, p ≤ 0.001, and negatively associated with anticipatory pleasure, r = − 0.39, p ≤ 0.001, and consummatory pleasure, r = − 0.32, p = 0.01, indicating that individuals with greater ARFID severity experience greater depression and lower pleasure. Age was not significantly related to anticipatory pleasure, r = − 0.09, p = 0.47, or consummatory pleasure, r = 0.19, p = 0.11.

Table 3 Correlations between study variables in the ARFID sample

The full results of regression models used to assess whether correlations remained significant when controlling for relevant covariates are presented in Table 5. The overall model testing the relationships between anticipatory pleasure, ARFID severity, depression, and BMI was significant, R2 = 0.19, F(3, 65) = 6.22, p < 0.001. Consistent with our hypothesis, ARFID severity was negatively related to anticipatory pleasure, B = − 0.30, p = 0.03. Depression, but not BMI, was significantly negatively associated with anticipatory pleasure, B = − 0.02, p = 0.03. These results indicate that, among individuals with ARFID, those with greater ARFID severity and greater depression severity report lower anticipatory pleasure.

The overall model testing relationships between consummatory pleasure, ARFID severity, depression, and BMI was also significant, R2 = 0.07, F(3, 65) = 2.80, p = 0.05. ARFID severity was the only variable in the model that was significantly related to consummatory pleasure, B = − 0.33, p = 0.04, which suggests individuals with more severe ARFID symptoms report lower consummatory pleasure independent of depression symptoms and BMI.

TEPS subscales without food items

The correlation coefficients between ARFID severity, anticipatory pleasure, and consummatory pleasure are presented in Table 4. ARFID severity was negatively associated with anticipatory pleasure, r = − 0.40, p ≤ 0.001, but its relationship with consummatory pleasure was no longer significant after omitting food-related TEPS items, r = − 0.16, p = 0.18. Similar to the results of analyses with the full TEPS subscales, age was not significantly related to anticipatory pleasure, r = 0.01, p = 0.93, or consummatory pleasure, r = 0.07, p = 0.59.

Table 4 Correlations between study variables and anticipatory and consummatory pleasure with food-related items omitted

The full results of regression models used to assess whether correlations remained significant when controlling for relevant covariates are presented in Table 5. When omitting food-related items, the overall model testing the relationships between anticipatory pleasure, ARFID severity, depression, and BMI remained significant, R2 = 0.17, F(3, 65) = 5.69, p = 0.002. ARFID severity remained negatively related to anticipatory pleasure, B = − 0.36, p = 0.02, but depression was no longer significantly associated with anticipatory pleasure, B = − 0.02, p = 0.06.

Table 5 Regression analyses exploring associations between pleasure and ARFID severity within the ARFID sample

In contrast to the model with the all TEPS items, the overall model testing relationships between consummatory pleasure, ARFID severity, depression, and BMI was not significant, R2 = 0.05, F(3, 65) = 2.21, p = 0.09.

Exploratory analyses of relationships between pleasure and ARFID profilesFull TEPS subscales

Scores on the sensory sensitivity profile were negatively associated with anticipatory pleasure, r = − 0.26, p = 0.03, as were scores on the lack of interest in food profile, r = − 0.31, p = 0.01; consummatory pleasure was not associated with any ARFID profile (see Table 3). These results suggest that, while individuals with greater sensory sensitivity and less interest in food may report lower anticipatory pleasure, none of the ARFID profiles were associated with level of consummatory pleasure.

The full results of a follow-up regression model controlling for depression and BMI are presented in Table 6. The overall model assessing relationships between ARFID profiles, depression, BMI, and anticipatory pleasure was significant, R2 = 0.17, F(5, 63) = 3.87, p = 0.04. Although sensory sensitivity and lack of interest in food were negatively related to anticipatory pleasure in correlation analyses, depression was the only variable in this model that had a significant relationship with anticipatory pleasure (B = − 0.03, p = 0.02). These results indicate that, among individuals with ARFID, greater depressive symptoms were associated with decreased anticipatory pleasure, regardless of specific ARFID profile.

Table 6 Regression analyses exploring associations between anticipatory pleasure and ARFID profiles within the ARFID sampleTEPS subscales without food items

Correlations between TEPS subscales and ARFID profiles are presented in Table 4. When food items were removed, the lack of interest in food profile remained negatively correlated with consummatory pleasure, r = − 0.33, p = 0.005, but the sensory sensitivity profile was no longer significantly associated with anticipatory pleasure. In contrast to analyses with all TEPS items, the fear of aversive consequences profile was positively correlated with consummatory pleasure when food items were removed, r = 0.27, p = 0.02. These results suggest that, when pleasure related to food is not assessed, individuals with the lack of interest ARFID profile experience decreased anticipatory pleasure and individuals with the fear of aversive consequences profile experience greater consummatory pleasure.

Because both anticipatory and consummatory pleasure were correlated with at least one subscale when food items were removed, we completed follow-up regression models for both TEPS subscales. The full results of regression models controlling for depression and BMI are presented in Table 6. The overall model assessing relationships between ARFID profiles, depression, BMI, and anticipatory pleasure remained significant, R2 = 0.15, F(5, 63) = 3.41, p = 0.01. Depression remained significantly negatively related to anticipatory pleasure, B = − 0.03, p = 0.05; in contrast with the model using all TEPS items, lack of interest score was significantly negatively associated with anticipatory pleasure, B = − 0.17, p = 0.05, suggesting that the relationship between these two constructs is strengthened when food-related pleasure is not measured.

The model assessing relationships between ARFID profiles, depression, BMI, and consummatory pleasure was not significant, R2 = 0.09, F(5, 63) = 2.27, p = 0.06, but depression was negatively related to consummatory pleasure, B = − 0.03, p = 0.03, indicating that depressive symptoms, rather than specific ARFID presentations, may contribute to the variance in consummatory pleasure within this population.