ولع مصرف دولت، در دسترس بودن مواد غذایی و واکنش پذیری برای غذاهای سرپایی ارجح

Abstract The startle response has been shown to be useful in studying reactivity to food cues. Following 6 h of food deprivation and exposure to neutral and food cues, we examined the role of state craving combined with both a short and long delay of consumption on affect and startle reflex. Participants completed the PANAS, consumed a controlled early morning meal, and experienced 6 h of food deprivation. They then reported back to the laboratory, completed a second baseline PANAS, and had their baseline eyeblink EMG startle responses to 100 dB(A) startle probe assessed. Prior to and following the presentation of cues, startle probes were presented and responses were recorded. The PANAS and state craving were also assessed after each cue. Food cues provoked higher levels of state craving than neutral cues and startle responses failed to habituate as quickly to food cues as they did to neutral cues. In addition, cue exposure created the highest NA among high state cravers in the long delay of consumption group. Startle responses differed from NA in that with long delay startle was high irrespective of state craving scores; in the short delay of consumption condition, startle increased linearly with state craving. These results illustrate that state craving and expectations of food availability are important variables in understanding food-related cue reactivity.

Results The two delay of consumption groups (short or long delay) had nearly identical mean (SD) BMI scores, 24.72 (3.52) vs. 24.02 (3.09), and BMI did not have a significant relationship with state craving. In addition, mean (SD) state craving scores were nearly identical for men and women: 3.43 (0.63) vs. 3.56 (0.56), respectively. Analysis of state craving and affect data Initial analyses were conducted to examine the state craving and affect data for the two cue conditions by the delay of consumption manipulation. In a cue type (tape vs. food) by delay of consumption (short vs. long) repeated measures general linear model for state craving, the only significant effect was for cue type, F(1,44) = 12.51, p < .01, partial η2 = 0.22. As shown in Table 1, the food cue produced higher state craving responses than did the neutral cue. As expected state craving scores for the two cues were highly related to one another, r = .80, p < .01. In addition, it is interesting to note that state craving scores were directed related to both early morning negative affect (NA), r = .32, p < .05, and baseline NA r = .40, p < .01. Table 1. Means ± SE for state craving by cue condition and delay of consumption. Cue type Delay manipulation Row means Short Long Neutral 3.32 ± 0.12 3.29 ± 0.11 3.30 ± 0.08 Food 3.50 ± 0.13 3.49 ± 0.12 3.49 ± 0.08 Column means 3.40 ± 0.12 3.39 ± 0.11 Table options Table 2 provides the raw data for the PANAS scores. Analyses of these data occurred in 2 steps. First, time of assessment (morning, baseline, neutral cue, and food cue) by delay of consumption (short vs. long) repeated measures general linear models were conducted to examine whether NA and positive affect (PA) changed across the course of the study and whether these effects were influenced by the delay manipulation. Inspection of the data in Table 2 suggests that NA increased from early morning to baseline and again from baseline to cue exposure. There were trends for NA to be higher in the long delay than short delay condition and no obvious patterns for positive affect (PA). In repeated measures analyses for NA and PA with time of assessment and delay as terms in the models, the only statistically significant effect was a main effect for NA—F(2,132) = 3.90, p < .05, partial η2 = 0.22—illustrating that morning NA (mean ± SE = 6.33 ± 0.24) was lower (p < .05) than NA at baseline (6.90 ± 0.31), and following either the neutral cue (7.32 ± 0.39), or the food cue (7.43 ± 0.39). Table 2. Raw means ± SE for affect by delay of consumption. PANAS scores Negative affect Positive affect Short delay Long delay Short delay Long delay Early morning 6.61 ± 0.36 6.04 ± 0.34 10.29 ± 0.79 10.44 ± 0.49 Baseline 6.76 ± 0.46 7.04 ± 0.43 10.95 ± 0.77 11.00 ± 0.71 Neutral cue 7.05 ± 0.57 7.60 ± 0.53 9.62 ± 0.79 9.72 ± 0.73 Food cue 6.90 ± 0.58 7.96 ± 0.53 11.42 ± 0.82 10.52 ± 0.76 Table options The second step was the primary analysis for NA. In this general linear model, the data for cue type was combined since there was no evidence that NA differed as a function of exposure to either the tape or food cue. Terms in the model included a main effect for delay of consumption, a main effect for state craving (combining data from both cues), the interaction of delay by state craving, and the use of baseline as a covariate. The model produced a statistically significant interaction term for delay by state craving, F(1,41) = 10.79, p < .01, partial η2 = 0.21. To assist with interpretation of this effect, we conducted a 3-way split on the state craving scores and plotted the interaction effect for state craving by delay (see Fig. 1). The highest NA occurred among participants who scored high on state craving and were exposed to a long delay of consumption. Simple effect correlational analyses between state craving and NA within the short and long delay conditions confirmed the interpretation of this effect; that is, there was no relationship (or a zero slope) between state craving and NA in the short delay condition (r = −.11, p > .05), yet a significant direct relationship between state craving and NA in the long delay condition (r = .66, p > .01). The interaction of the fasting manupulation and state craving on negative ... Fig. 1. The interaction of the fasting manupulation and state craving on negative effect. Figure options Startle magnitude The baseline startle data, collected before the delay manipulation and exposure to the experimental cues, varied somewhat between the four groups of participants (see Table 3). These differences were not statistically significant (p for cue order = .554, for delay of consumption p = .097, and for the cue order by delay of consumption interaction p = .150) and could not have been due to either cue order or delay of consumption, since group assignment was completed and baseline startle data were measured before either variable was introduced. However, these random between group differences at baseline may have obscured our ability to evaluate the hypotheses of interest. Thus, to decrease the impact of this inter-participant variability on our analyses, startle magnitude was analyzed as a ratio of post-cue responses divided by pre-cue baseline values. Table 3. Means ± SE for startle data by delay of consumption. Assessment Short delay Long delay Baseline (microvolts) 2.51 ± 0.57 3.65 ± 0.52 Neutral cue (microvolts) 2.06 ± 0.58 3.49 ± 0.53 Food cue (microvolts) 2.07 ± 0.52 3.49 ± 0.58 Neutral cue (ratio scale) 0.83 ± 0.07 0.96 ± 0.07 Food cue (ratio scale) 0.82 ± 0.08 1.10 ± 0.08 Table options The first analysis was a repeated measures general linear model examining the effects of cue order, cue type, and delay of consumption on startle magnitude. A significant main effect for delay of consumption was found, F(1,42) = 5.534, p < .025, partial η2 = 0.17, with larger proportional startle values for long than short delay (mean ± SE ratios of 1.04 ± 0.06 and .83 ± 0.06, respectively). In addition, there was a significant interaction for cue type by cue order, F(1,42) = 4.91, p < .05, partial η2 = 0.10 (see Fig. 2). Simple effect testing within each cue order revealed that startle magnitude was higher for the food cue than for the neutral cue when the neutral cue was presented first (p < .05), however, when food was presented in the 2nd position, the startle response for the neutral and food cues were essentially identical (p = .33). In essence, the presence of food in the 2nd position blunted the habituation effect that was observed when the neutral cue appeared 2nd (see Schicatano & Blumenthal, 1994, for a similar finding with caffeine). The interaction of cue type and cue order. Fig. 2. The interaction of cue type and cue order. Figure options In a final analysis, state craving and the interaction of state craving by delay were included as additional effects. In addition to the previously discussed findings, this model yielded a statistically significant state craving by delay interaction term, F(1,40) = 4.42, p < .05, partial η2 = 0.10. To elucidate the meaning of this interaction, Fig. 3 provides the startle magnitude data partitioned by delay and a 3-way split on state craving scores. These data illustrate that startle magnitude increased with state carving scores within the short delay condition, yet were unrelated to state craving in the long delay condition; that is, startle responses were elevated in the long delay condition irrespective of the level of state craving. Simple effect correlational analyses between state craving and startle magnitude within the short and long delay conditions confirmed the interpretation of this effect; that is, there was a significant relationship between startle magnitude and state craving within the short delay condition (r = .61, p < .01), yet no relationship (or a zero slope) between state craving and startle magnitude within the long delay condition (r = −.06, p > .05). The interaction of state craving with the delay manipulation on startle ... Fig. 3. The interaction of state craving with the delay manipulation on startle magnitude.