Neurophysiological Effects Associated With Subliminal Conditioning of Appetite Motivations

Abstract

When attempting to encourage eating, explicitly providing statements like "eating is pleasant" may produce little effect. This may be due to subjective, negatively-valenced narratives evoked by perception of the verb "eating" (e.g., eating →fat →lonely), overriding any explicitly provided eating-pleasant valence information. In our study, we presented eating-related verbs under subliminal visual conditions to mitigate the onset of eating-associated deliberation. Verbs were linked with neutral or positively valenced terms across independent blocks. Modulations of event-related magnetoencephalographic (MEG) components and parietal activations in the alpha range (8-12 Hz) illustrated a significant effect of valence during pre-lexical time windows. We found significantly greater saliva production and declarations of increasing hunger after eating-related verbs were linked with positive terms. Orally reported preferences did not vary between conditions.

Keywords: appetite; eating disorders; learning theory; magnetoencephalography; motivation; subliminal conditioning.

FIGURE 1

Valences predicated through the proposition “eating is pleasant” can be countered by subjective narratives evoked by conscious projection of the word “eating” (dashed arrows, top panel). Subjective and explicit valences are assumed to integrate within 200–300 ms of word onsets (before lexical processing). We aimed to minimize subjective eating-associated narratives by presenting “eating”-related words subliminally (bottom panel).

FIGURE 2

Each conditioning trial commenced with a fixation cross on the left or right sides of the screen. Subjects had to respond to the location of the cross by pressing “1” or “2” on a button box. This produced a CS sandwiched between masks, a blank interval and a US in the same location as the cross. US offsets produced a second blank interval, followed by two two-alternative forced choice tasks measuring CS visibility and motivational state, respectively.

FIGURE 3

(A) Grand averaged activity over left central and parietal clusters over a 2400 ms window (left) and corresponding source activations during neutral (middle) and positive (right) blocks. CS and US onsets were at 0 and 1170 ms, respectively. Early negative (N1) and positive (P1) event-related components produced by CS and US onsets differed significantly between neutral and positive conditions (asterisks indicate p’s < 0.02) across the displayed clusters. (B) Mean activation maps across five 1-Hz bands corresponding to the alpha range. Alpha power did not vary significantly between left and right regions, or between stimulus types. Significantly greater alpha (p = 0.001) appeared during positive conditions only. Error bars indicate SDs.

FIGURE 4

Subjects produced significantly more saliva after positive conditioning blocks (A). Subjects responded Yes significantly more times (B) and faster (C) in response to the question “Are you getting hungry?” during positive conditioning trials. Oral preferences did not differ between conditions for CS (D) or distractors (not shown), where higher values indicate increased preference (y-axis). CS were identified significantly more times during positive conditioning trials (E), although hit rates across both conditions fell below chance levels. Response times for visibility checks did not vary across conditions (F). Asterisk indicate p’s < 0.03.