TOPLINE:

Smoking behaviour in individuals at clinical high risk for psychosis (CHR-P) was not associated with altered functional connectivity in major brain networks. A large longitudinal study of 486 participants at CHR-P, followed up over 8 months, found no significant differences in the within- or between-network connectivity of the default mode network (DMN), salience network (SN), or frontoparietal networks (FPNs) between smoking and non-smoking groups.

METHODOLOGY:

• Researchers analysed resting-state functional MRI data from the NAPLS-3 conducted from 2015 to 2020.
• A total of 486 participants aged 12-30 years who met the criteria for CHR-P on the basis of the structured interview for psychosis risk syndromes and had at least one MRI scan available were included at baseline.
• Of the 486 participants, 101 (21%) were smokers and 385 (79%) were non‑smokers. Among smokers, 58% smoked occasionally, 22% smoked less than 10 cigarettes per day, 18% smoked 10-25 cigarettes per day, and 2% smoked more than 25 cigarettes per day.
• A total of 1128 scans were collected across baseline and 2-, 4-, 6-, and 8-month follow-up visits; functional connectivity for the DMN, SN, and left and right FPNs was extracted, and within- and between-network strengths were assessed.
• Statistical models evaluated longitudinal connectivity changes between smoking and non-smoking CHR-P groups, with covariates including age, sex, cannabis use, alcohol use, and MRI site, using false discovery rate (FDR) correction for multiple comparisons.

TAKEAWAY:

• At baseline, participants at CHR-P in the smoking group showed stronger connectivity within the SN than those in the non-smoking group (P = .041), but this finding did not survive FDR correction (P_FDR = .164).
• At baseline, participants at CHR-P in the smoking group demonstrated stronger connectivity between the left and right FPNs (P = .037); however, this finding did not survive multiple comparison correction (P_FDR = .184).
• Over the 8-month follow-up period, longitudinal analyses revealed no significant group or time-by-group effects on within-network or between-network functional connectivity between participants at CHR-P in smoking and non-smoking groups.
• Participants at CHR‐P in the smoking group were significantly older, were more often male, had more years of education, used cannabis more frequently and heavily, and consumed alcohol more often and in greater quantities (P < .001).

IN PRACTICE:

"Our findings do not support that mild smoking is associated with baseline or eight‐month follow‐up within‐ and between‐network functional connectivity of the DMN, SN and FPN in CHR‐P individuals. This suggests that smoking is not associated with large‐scale network changes at this early stage of illness," the authors wrote.

"However, longitudinal studies with higher smoking intensity and longer follow‐up than eight months are needed to clarify whether network alterations emerge with heavier and more long-standing use or illness progression," they added.

SOURCE:

This study was led by Merel Koster, University of Amsterdam, Amsterdam, Netherlands. It was published online on March 11, 2026, in General Psychiatry.

LIMITATIONS:

The study lacked detailed information on smoking behaviour, particularly regarding the age of smoking initiation and the timing of the last cigarette prior to scanning, which prevented accounting for acute nicotine effects or withdrawal. The study sample primarily consisted of light smokers with limited nicotine exposure, which may have been insufficient to cause detectable connectivity changes. The study also lacked data on the timing of the last cigarette before scanning and cumulative pack-years. The study did not control for the use of antipsychotic medications. The sample likely included primarily high-functioning individuals who were able to provide consent and complete the demanding longitudinal MRI study, which could have introduced selection bias.

DISCLOSURES:

The authors reported that they did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors for this project and declared having no conflicts of interest.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.