Each participant will attend three sessions where the participant will experience three distinct cue-induced craving tasks in a randomized fashion.

For the neutral cue condition, the participant will be seated in front of a 16-inch monitor and instructed to sit upright to view 20 pictures of water bottles via E-Prime software (Psychology Software Tools Inc., PA), presented for five seconds each in randomized order. The images will include pictures of a water bottle alone, water being poured from a bottle to a glass, and an individual drinking from a bottle of water. Next, participants will watch as research staff open a bottle of water, pour it into a glass, and place it on a table in front of the participant. Participants will also be asked to look at a water bottle, hold it, sniff it, and take a drink of water.

Immediately following each cue condition, participants will be asked the following open-ended questions: (1) Describe your mood, (2) Describe how you feel, and (3) What symptoms are you experiencing? Responses will be recorded using a digital audio recorder.

Qualitative data collected on the symptoms, moods and feelings following cue-induced craving tasks will be transcribed and independently coded for common craving domains based on our published preliminary study by two raters. Raters will also document other potential domains that do not fit into any of the a priori categories. Ad Hoc domains require that both raters agree that a new category should be created. Independently coded responses will be compared and incongruencies will be reviewed by a third rater who will make the final designation. The types and frequencies of craving domains will be summarized with descriptive statistics.

Responses to the question "What was this question designed to measure?" will be coded by two independent raters as either "Demonstrates understanding" or "Does not demonstrate understanding". Next, responses to the question "What would you change, if anything, about this measure?", will be transcribed and independently coded by two raters as either "no changes needed" or "changes needed". Responses will be compared and incongruencies will be reviewed by a third rater who will make the final designation. If final responses to either question is determined to be either "does not demonstrate understanding" and/or "changes needed", the question will be labeled as unacceptable. For each craving assessment item, the proportion of participants who describe the item as unacceptable will be computed. The proportions of responses designated as "does not demonstrate understanding" and "changes needed" will also be computed.

Responses to the question "What was not captured by these questions that you would include in a measure for assessing opioid craving?" will be coded for themes by two independent raters. Independently coded responses will be compared and incongruencies will be reviewed by a third rater who will make the final designation. The types and frequencies of recommendations will be summarized with descriptive statistics.

Average measures of heart rate and galvanic skin response 10 minutes prior to cue-exposure will be subtracted from average measures of heart rate and galvanic skin response for the 10 minutes beginning with cue-exposure to calculate a change from baseline measure of both physiological measures. Greater changes will be indicative of a greater conditioned response. (1) Logistic mixed effects models will evaluate the association between change from baseline physiological measures and whether an individual described a given craving domain across craving condition (domain present, domain absent) and stakeholder group . (2) Linear mixed effects models will evaluate the association between change from baseline heart rate and galvanic skin response and scores on existing craving question single item scores (range: 0-100 and 0-9) across craving condition and stakeholder group. Greater scores on single item craving questions are associated with greater craving.