Preliminary investigation of between-network connectivity and craving during early alcohol abstinence

This Level 3 Registered Report investigates whether deviations in functional connectivity between the salience, default mode, and central executive networks, measured via resting-state fMRI and normative models, underlie craving during early alcohol abstinence in individuals with alcohol use disorder.

The Gist

Imagine your brain is a bustling city with three major departments that need to work together to keep you functioning smoothly:

1. The "Daydreaming" Department (Default Mode Network): This is where your mind wanders when you're bored. It handles your memories, your sense of self, and your inner monologue.
2. The "Boss" Department (Central Executive Network): This is your focus team. It handles planning, problem-solving, and making tough decisions when you need to get work done.
3. The "Alarm Clock" Department (Salience Network): This is the switchboard operator. Its job is to scan the world and your body, decide what is important right now, and tell the other two departments to either wake up or stand down.

The Problem: A Broken Switchboard
For people with Alcohol Use Disorder (AUD), this city is in chaos. When they try to stop drinking, their "Alarm Clock" gets confused. Instead of noticing that it's time to eat lunch or talk to a friend, it gets hyper-focused on the urge to drink. It rings the alarm bell so loudly that the "Daydreaming" department gets stuck in a loop, obsessing over alcohol, while the "Boss" department struggles to take control.

The Study: Checking the City's Wiring
This research paper is like a team of city engineers trying to fix the wiring in this brain-city during the very first few weeks of quitting alcohol.

• The Goal: They want to see if the strength of the connection between the "Alarm Clock" and the "Daydreaming" department predicts how badly a person craves alcohol.
• The Participants: They studied 27 people who were just starting their journey to sobriety (in a hospital detox program) and 17 healthy people who drink socially but don't have a disorder.
• The Timeline: They scanned the brains of the recovering group twice:
  1. Day 1: Right when they started detox (when cravings are usually at their peak and the body is shaking).
  2. Day 18: After two weeks of being sober and receiving therapy.
• The "Craving" Meter: They asked the patients how much they were obsessing about drinking using a questionnaire called the OCDS.

The Secret Weapon: The "Normative Model"
Here is the clever part. Instead of just comparing the "sick" brains to the "healthy" brains (which is like comparing a broken car to a new one), the researchers used a massive digital blueprint of what a "normal" brain looks like at every age.

Think of it like a tailor. Instead of just saying "this suit is too big," they compare the patient's brain to a giant database of thousands of healthy people of the exact same age. This allows them to say, "This specific person's brain wiring is 2 standard deviations off the norm for their age." This helps them see the true damage caused by alcohol, separate from normal aging.

What They Plan to Find Out
The researchers have a specific theory they are testing:

• The Hypothesis: They believe that when a person is in early withdrawal, their "Alarm Clock" (Salience) is stuck in a high-alert mode, constantly ringing the bell for the "Daydreaming" department to think about alcohol. They expect to see a super-strong, messy connection between these two networks.
• The Prediction: They think that the stronger this messy connection is on Day 1, the more the person will crave alcohol.
• The Hope: They also want to see if, by Day 18, as the person gets sober and the cravings go down, this "messy connection" starts to calm down and return to normal.

Why This Matters
If they are right, this study could prove that craving isn't just a "weakness of will." It's a physical glitch in the brain's wiring.

If we can identify exactly which wires are crossed, doctors might be able to develop better treatments—like specific therapies or medications—that target just those connections to help people quit drinking and stay sober. It's about fixing the city's wiring so the "Boss" department can take back control from the "Alarm Clock."

In a Nutshell:
This paper is a "pre-registered" investigation (meaning they wrote down their plan before looking at the results to ensure honesty) to see if the way different parts of the brain talk to each other explains why alcohol cravings are so powerful, and if fixing that conversation helps people recover.

Technical Summary

Based on the provided manuscript, here is a detailed technical summary of the study.

1. Problem Statement

Alcohol Use Disorder (AUD) is a chronic condition with high relapse rates, particularly within the first three months of abstinence. Craving during early withdrawal is a primary predictor of relapse, yet the specific neurobiological mechanisms driving this craving remain unclear.

• Gap in Literature: While previous studies have examined alcohol cue reactivity and resting-state functional connectivity (rsFC) in AUD, few have investigated patients during acute withdrawal due to challenges like motion artifacts and benzodiazepine administration. Furthermore, existing studies often rely on simple case-control comparisons that fail to account for individual heterogeneity and age-related brain changes.
• Theoretical Framework: The study is guided by Menon's Triple Network Model (TNM), which posits that psychopathology arises from dysregulation in the interactions between the Salience Network (SN), Default Mode Network (DMN), and Central Executive Network (CEN). The authors hypothesize that excessive connectivity between the SN and DMN may underlie the internal focus and heightened salience of craving during acute withdrawal.

2. Methodology

This study is a Level 3 Registered Report (Peer Community In), meaning data collection and preprocessing were completed prior to hypothesis testing, and key subject-level variables have not yet been observed by the authors.

• Participants:
  • AUD Group: 27 individuals recruited from an inpatient withdrawal program (Cliniques Universitaires Saint-Luc, Belgium). They underwent two MRI sessions: Day 1 (acute withdrawal, 6–8 hours post-admission) and Day 18 (early abstinence).
  • Control Group: 17 healthy controls scanned twice, two weeks apart, to adapt the normative model to the local site.
  • Exclusion: Severe psychiatric comorbidities, other substance abuse (except nicotine), and excessive motion artifacts.
• Clinical Measures:
  • Craving: Assessed using the Obsessive Compulsive Drinking Scale (OCDS), specifically the "obsessive thoughts" subscale (items 1–6).
  • Withdrawal Severity: Measured via the Cushman score (physiological parameters).
  • Medication: Patients received a tapering regimen of oral diazepam and thiamine.
• Imaging Protocol:
  • Scanner: 3T GE SIGNA Premier.
  • Sequence: Resting-state fMRI (eyes closed, 217 volumes, TR=1700ms) and structural T1.
  • Preprocessing: Performed using fMRIPrep (v20.2.7) and CONN toolbox. Steps included motion correction, susceptibility distortion correction, normalization to MNI space, and denoising (regressing out white matter, CSF, motion parameters, and using aCompCor).
• Normative Modeling Approach (Key Innovation):
  • Instead of raw connectivity values, the study utilizes lifespan normative models derived from a large reference cohort ($N=21,515$) via the PCNportal.
  • Individual connectivity matrices (Smith-10 parcellation) are compared against age-matched healthy distributions to generate Z-score deviation scores. This accounts for aging effects and site-specific variations more rigorously than standard covariate adjustment.
• Planned Statistical Analysis:
  • Q1 (Cross-sectional): Correlation between DMN–SN connectivity Z-scores and craving scores at Day 1 (AUD group only).
  • Q2 (Longitudinal): Correlation between the change in DMN–SN connectivity ($\Delta Z$) and the change in craving ($\Delta$ OCDS) between Day 1 and Day 18.
  • Exploratory: Analysis of other TNM pairs (SN–CEN, DMN–CEN) with False Discovery Rate (FDR) correction.
  • Power: The study is powered (90%) to detect large effect sizes ($r \geq 0.52$) given the sample size ($N=27$).

3. Key Contributions

• Novel Timing: This is one of the few studies to examine rsFC during acute alcohol withdrawal (Day 1), a critical period often skipped due to methodological difficulties.
• Methodological Rigor: The use of normative modeling allows for the quantification of subject-specific deviations from typical brain patterns, effectively disentangling disease effects from normal aging and site variability.
• Registered Report Design: By pre-registering the analysis plan before observing the outcome variables, the study minimizes bias and p-hacking, adhering to high standards of reproducibility.
• Theoretical Application: It directly applies the Triple Network Model to the specific mechanism of craving in AUD, moving beyond general network dysfunction to specific inter-network connectivity hypotheses.

4. Results

Note: As this is a Stage 1 manuscript (pre-results), no empirical results or statistical findings are reported in this document.

• The manuscript outlines the proposed analyses and the rationale for them.
• The authors explicitly state that subject-level values for the primary variables (IDP Z-scores and craving scores) have not been inspected or observed.
• The document serves as a protocol for how the data will be analyzed to test the hypothesis that higher craving correlates with stronger DMN–SN connectivity.

5. Significance

• Clinical Implications: If the hypothesis is supported, identifying DMN–SN hyperconnectivity as a marker of craving could lead to personalized interventions (e.g., neuromodulation or pharmacotherapy) targeting specific network dysfunctions to prevent relapse.
• Scientific Advancement: The study advances the understanding of the neurobiological phenotypes of AUD by moving from static group differences to dynamic, individualized deviations from normative brain function.
• Relapse Prediction: By focusing on the acute withdrawal phase, the findings could help identify high-risk patients early in the recovery process, allowing for more targeted and timely therapeutic interventions.

In summary, this paper presents a rigorously designed, pre-registered protocol to investigate the neural basis of alcohol craving using advanced normative modeling. While results are pending, the study promises to provide high-quality evidence regarding the role of Triple Network dysconnectivity in early AUD recovery.