How Often People Use Substances May Reveal a Wider Web of Mental Health Risk

Frequent use of alcohol, cannabis, or stimulants has long been linked to strange and unsettling experiences: hearing sounds that are not quite right, feeling flooded by noise, or holding beliefs that others find hard to follow. A new study published in the International Journal of Mental Health and Addiction suggests that these psychotic-like experiences rarely travel alone. Instead, they appear to sit inside a dense web of psychological vulnerabilities, and the tighter that web becomes, the more frequently a person reports using substances. The research, led by Maria Nalberczak-Skóra of the Polish Academy of Sciences, offers one of the most detailed maps yet of how substance use, unusual experiences, and everyday mental health difficulties are wired together in the general population.

Psychotic-like experiences, often abbreviated as PLEs, are subclinical phenomena that fall short of a diagnosable psychotic disorder. They include delusion-like experiences such as paranoia, suspiciousness, or the feeling that one’s thoughts are being broadcast, as well as perceptual abnormalities, most commonly in the auditory domain, where sounds seem distorted, intensified, or overwhelming. These experiences are surprisingly common: roughly 5 to 10 percent of the general population report them at any given time, and more than a quarter of people endorse at least one such experience over their lifetime. Although most are transient, about 20 percent of people who report acute psychotic-like experiences go on to develop persistent symptoms, which is why researchers treat them as important early warning signals of broader psychopathology.

The connection between substance use and these experiences has been established in previous work. A systematic review and meta-analysis of young people found that substance users were nearly twice as likely to report psychotic-like experiences as nonusers. High-potency cannabis has been linked to more frequent and severe paranoia, alcohol use has been implicated in psychotic symptoms particularly among people with mood disorders, and high doses of methamphetamine can induce psychosis even in otherwise healthy individuals. What has remained unclear is how these associations vary with how often someone uses substances, and how they are shaped by the wider psychological landscape, including impulsivity, recent stress, childhood adversity, and other psychiatric symptoms.

To untangle this, the researchers recruited 1,196 adults aged 18 to 51 through online advertisements and collected anonymous self-report data between July 2024 and January 2026. Rather than examining each risk factor in isolation, they used network analysis, a statistical framework in which every variable is represented as a node and every unique association between variables as an edge, with all other variables statistically accounted for. The network included frequency of alcohol, cannabis, and stimulant use, impulsivity, other psychopathological symptoms such as depression and anxiety, recent stress exposure, adverse childhood experiences, delusion-like experiences, and auditory perceptual abnormalities. This approach allowed the team to ask whether substance use connects directly to psychotic-like experiences or whether the connection runs indirectly through the surrounding psychological terrain.

A key innovation was the way participants were grouped. Instead of imposing an arbitrary cutoff, the researchers used a data-driven network tree procedure that examined how the correlation structure among the psychological variables changed with substance use frequency. This analysis identified a boundary between use at weekly-or-less and use more than weekly, splitting the sample into nonusers (308 people), lower frequency users (596), and higher frequency users (292). The authors stress that this boundary is exploratory and sample-derived rather than a clinical threshold, and sensitivity analyses showed that while the decision to split the sample was highly consistent, the exact location of the boundary varied somewhat across resampling procedures and substances.

The results were striking. In the full-sample network, the three substance use variables formed the strongest connections with one another, particularly between alcohol and cannabis. But the most important finding concerned the wider architecture: higher frequency users showed significantly greater connectivity within the psychosocial and psychotic-like experience network than nonusers, even though the overall structure of that network was similar. In other words, frequent users did not display a qualitatively different pattern of psychological difficulties, but the links among their symptoms, stress, adversity, impulsivity, and unusual experiences were measurably stronger. The full nine-node networks of lower and higher frequency users also differed in both global strength and overall structure, although the only individual edges that survived statistical correction were those among the substance use variables themselves.

Delusion-like experiences emerged as the star of the show. In the full network, delusion-like experiences showed direct conditional associations with all three substances, whereas auditory perceptual abnormalities showed no direct connection to any of them. Delusion-like experiences also carried the highest bridge strength, a measure of how strongly a node connects different domains of the network, followed by other psychopathological symptoms. Among higher frequency users, all three substances showed statistically supported total associations with delusion-like experiences, while none of the substances showed a supported total association with auditory perceptual abnormalities. This suggests that unusual beliefs and suspicious interpretations may be more closely tied to frequent substance use than the perceptual distortions assessed here, and it supports the growing view that different types of psychotic-like experiences should not be lumped together as a single construct.

The exploratory mediation analyses added another layer of nuance. Among lower frequency users, supported total associations with psychotic-like experiences were confined to stimulant use, and they involved impulsivity, other psychopathological symptoms, and adverse childhood experiences. Among higher frequency users, the picture shifted: associations with delusion-like experiences for all three substances were carried most consistently by other psychopathological symptoms and recent stress exposure, with no supported direct associations remaining. In the full sample, indirect components accounted for large portions of the total associations, sometimes more than 70 percent. The authors are careful to emphasize that because the study is cross-sectional, these mediation models describe statistical decompositions of concurrent associations, not causal or temporal pathways.

The clinical implications are considerable. If frequent substance use marks a broader mental health burden rather than an isolated behavioral exposure, then assessing someone who drinks or uses cannabis daily should arguably extend beyond substance frequency alone to include psychiatric symptoms, stress, impulsivity, childhood adversity, and psychotic-like experiences. The finding that delusion-like experiences deserve particular attention among frequent users could help clinicians decide where to look first. At the same time, the authors caution against overinterpreting centrality metrics: a node with high bridge strength is not necessarily an intervention target, and changing one variable in a network does not guarantee changes elsewhere.

The study has clear limitations that the authors acknowledge candidly. Frequency of use is an incomplete measure of exposure, since it says nothing about quantity, potency, binge patterns, route of administration, or age at initiation. Group membership was defined by the highest frequency reported across the three substances, so the higher frequency group contained a mix of single-substance and polysubstance patterns, with polysubstance involvement especially common among frequent users. All measures were self-reported, and factors such as sleep, tobacco use, and family history of psychosis were not assessed. What the study does deliver is a compelling statistical portrait: frequent substance use is embedded in a more tightly interconnected psychological system, one in which delusional thinking, general psychiatric symptoms, stress, and adversity all pull on one another. Longitudinal research with richer measures of exposure will be needed to determine whether frequent use drives this entanglement, whether psychological distress drives frequent use, or whether both unfold from shared roots.

Subject of Research: Associations between substance use frequency and psychotic-like experiences within a broader psychosocial network

Article Title: Substance Use Frequency and Psychotic-Like Experiences Within a Broader Psychosocial Network

Article References: Nalberczak-Skóra, M., Bagrowska, P., Olechowska, A., & Gawęda, Ł. (2026). Substance Use Frequency and Psychotic-Like Experiences Within a Broader Psychosocial Network. International Journal of Mental Health and Addiction. https://doi.org/10.1007/s11469-026-01736-x

Image Credits: AI Generated

DOI: 10.1007/s11469-026-01736-x

Keywords: substance use, psychotic-like experiences, network analysis, delusion-like experiences, cannabis, alcohol, stimulants, impulsivity, adverse childhood experiences, stress, psychopathology, mental health