Neurobehavioral Affective Control Training

ERI will be measured using the self-rated Three-Factor Impulsivity Index "Feelings Trigger Action" scale using standard scoring methods (range: 1-5; larger values indicate higher levels of this characteristic). The researchers will perform a 2 (Arm/Condition: Experimental vs. Control) X 2 (Time: T1 vs. T2) Multivariate Analysis of Covariance (MANCOVA) to evaluate if N-ACT completion is associated with greater pre/post reductions in ERI scores compared to waitlist, as predicted.

Rumination will be measured using the self-rated "Brooding" subscale of the Ruminative Responses Scale using standard scoring methods (range: 5-20; larger values indicate higher levels of this characteristic). The researchers will perform a 2 (Arm/Condition: Experimental vs. Control) X 2 (Time: T1 vs. T2) MANCOVA to evaluate if N-ACT completion is associated with greater pre/post reductions in Rumination scores compared to waitlist, as predicted.

Valence-specific indices of stop-signal reaction time (SSRT) will be derived from the Emotional Stop-Signal Task using standard scoring procedures (measured in milliseconds, with larger values indicating less efficient emotional response inhibition, i.e., longer latency of relevant inhibitory processing). The researchers will perform a 2 (Arm/Condition: Experimental vs. Control) X 2 (Time: T1 vs. T2) MANCOVA to evaluate if N-ACT completion is associated with greater pre/post improvement in emotional response inhibition across valence (i.e., reductions in Positive SSRT and Negative SSRT, tested in the same model) compared to waitlist, as predicted, suggesting improvement in targeted aspects of affective control.

Valence-specific indices of working memory accuracy will be derived from the Memory and Affective Flexibility Task using standard scoring procedures (range: 0.00%-100%, measured as a percentage of working memory trials with correct responses over the total number of trials within each stimulus valence category, with larger values indicating more accurate working memory recall). The researchers will perform a 2 (Arm/Condition: Experimental vs. Control) X 2 (Time: T1 vs. T2) MANCOVA to evaluate if N-ACT completion is associated with greater pre/post improvement in emotional working memory across valence (i.e., increases in Positive Working Memory Accuracy and Negative Working Memory Accuracy, tested in the same model) compared to waitlist, as predicted, suggesting improvement in targeted aspects of affective control.

Valence-specific indices of accuracy (range: 0.00%-100%, measured as a percentage of "switch" trials with correct responses over the total number of trials within each stimulus valence category, with larger values indicating more accurate affective stimulus processing) and reaction time (RT; measured in milliseconds, with larger values indicating less efficient affective flexibility, i.e., longer latency of stimulus processing) will be derived from the Memory and Affective Flexibility Task using standard scoring procedures. A 2 (Arm/Condition: Experimental vs. Control) X 2 (Time: T1 vs. T2) MANCOVA will be used to evaluate if N-ACT completion is associated with greater pre/post improvement in affective flexibility across valence (i.e., increases in Accuracy and decreases in RT for Positive and Negative stimuli, tested in the same model) compared to waitlist, as predicted, suggesting near-transfer of trained abilities to non-targeted aspects of affective control.

Stop-signal reaction time (SSRT) will also be derived from the traditional (i.e., non-emotional) version of the Stop-Signal Task using standard scoring procedures (measured in milliseconds, with larger values indicating less efficient cold response inhibition, i.e., longer latency of relevant inhibitory processing). A 2 (Arm/Condition: Experimental vs. Control) X 2 (Time: T1 vs. T2) MANCOVA will be used to evaluate if N-ACT completion is associated with greater pre/post improvement in cold response inhibition (i.e., reductions in SSRT) compared to waitlist, as predicted, suggesting far-transfer of trained abilities to non-targeted aspects of cold cognitive control.

Cold working memory performance will be derived from the Backwards Digit Span task using standard scoring procedures (measured as mean digit span, i.e., the average number of correctly-recalled digits in reverse order from their presentation, with larger values indicating better cold working memory updating capacity). A 2 (Arm/Condition: Experimental vs. Control) X 2 (Time: T1 vs. T2) MANCOVA will be used to evaluate if N-ACT completion is associated with greater pre/post improvement in cold working memory (i.e., increases in mean digit span) compared to waitlist, as predicted, suggesting suggesting far-transfer of trained abilities to non-targeted aspects of cold cognitive control.

Internalizing psychopathology will be represented as a latent variable derived from raw scores on the self-rated (range: 1-5) Inventory of Depression and Anxiety Symptoms (IDAS) using confirmatory factor analysis (CFA), which is expected to produce a single-factor solution. CFA will be used to generate aggregated, standardized Z-scores on a latent composite "Internalizing" variable at each assessment point (coded such that larger values reflect greater symptom severity). Structural equation modeling of measurement invariance (between groups and within-subjects over time) will test hypothesized treatment effects on Internalizing psychopathology as a latent construct, which is expected to decrease in magnitude from pre- (T1) to post-intervention (T2).

Externalizing psychopathology will be represented as a latent variable derived from raw scores on self-rated scales pertaining to substance use (range: 1-5) from the revised Externalizing Symptom Inventory using confirmatory factor analysis (CFA), which is expected to produce a single-factor solution. CFA will be used to generate aggregated, standardized Z-scores on a latent composite "Externalizing" variable at each assessment point (coded such that larger values reflect greater symptom severity). Structural equation modeling of measurement invariance (between groups and within-subjects over time) will test hypothesized treatment effects on substance-related Externalizing psychopathology as a latent construct, which is expected to decrease in magnitude from pre- (T1) to post-intervention (T2).

Mean scores on the "Interest/Enjoyment" subscale of the Adapted Intrinsic Motivation Inventory will be used to evaluate the perceived acceptability of the N-ACT program. This self-report index comprises an average of seven items each rated on a 7-point Likert scale (range: 7-49), with larger values indicating greater subjective levels of treatment acceptability. Participants will complete these items during the final intervention session. An average rating of 35 (out of 49) points on this scale will be used as a quantitative "cutoff" score indicating hypothesized adequacy of the intervention's acceptability.

Adherence for each participant will be calculated as a function of (1) number of completed N-ACT training task trials divided by the total number of trials, i.e., individual differences in "dosage" of training and (2) attendance rates over the eight weekly N-ACT training sessions (range: 0.00%-100%, measured as percentages, with larger values reflecting higher levels of treatment adherence and feasibility). Median person-level adherence levels are expected to be between 70-75%, and are expected to decrease at the group-level by approximately 5% each week, with the majority of participants (i.e., > 66%) attending and completing all eight N-ACT training sessions.