We tested PPI and PPF of the eyeblink startle reflex to acoustic stimuli in participants at high clinical risk for psychosis and a matched control group. Both groups included a proportion of subjects who had recently used psychoactive substances, primarily cannabis, as confirmed by urinary drug screening on the day of testing. Because there was significant habituation across blocks we tested 1st block trials alone as well as all trials together. In the 1st block ARMS participants showed reduced PPI relative to controls, and this appeared to be driven by a group by cannabis use interaction, with recent use reducing PPI in ARMS participants but not in controls. When the analysis was limited to UDS-negative participants there was significantly reduced PPF in ARMS subjects relative to controls, but no differences in PPI. PPI and PPF both related to overall function in the ARMS as measured by the GAF.

A reduction in PPI in ARMS participants is consistent with the type of deficit evident in participants with psychotic disorders (Parwani et al., 2000; Wynn et al., 2004), their relatives (Kumari et al., 2005), and in people with psychosis-prone personalities (Kumari et al., 2008a). This suggests that PPI may represent a biomarker for vulnerability to psychosis, and adds weight to evidence of the validity of the Attenuated Psychosis Syndrome that has been much debated in DSM-V (Ruhrmann et al., 2010). However, our findings also suggest that cannabis use has differential effects on PPI in ARMS and controls; when cannabis users were excluded from the analyses PPI in the ARMS group was normal. In previous studies of PPI where urine testing for recent substance use was not conducted, including in samples of psychotic subjects (Parwani et al., 2000) and in clinical (Quednow et al., 2008), and genetic (Wynn et al., 2004) high risk samples, this raises the possibility that these findings may have been confounded by the effects of cannabis use. A further possibility is that those who were positive for cannabis on urinary testing are ‘true’ at risk for psychosis participants whose altered PPI is part of the prodrome of later psychosis. The present cross sectional data are unable to answer this interesting possibility.

A differential effect of cannabis on PPI in high risk subjects is in line with evidence of the symptomatic effects of cannabis in this group (Henquet et al., 2005; van Os et al., 2010). It is also consistent with evidence that effects of cannabis use on psychotic symptoms and on the risk of developing a psychotic disorder are moderated by genetic risk for psychosis and specific polymorphisms (Caspi et al., 2005; Henquet et al., 2008).

A differential effect of cannabis attenuating PPI in high risk subjects is also consistent with evidence that chronic cannabis users have reduced PPI, although only in attended-to trials (Kedzior and Martin-Iverson, 2006), and that drug free chronic cannabis users have PPI comparable to controls (Quednow et al., 2004). The finding of increased PPI in cannabis positive controls meanwhile is similar to another recent study (Preller et al., 2013) and consistent with evidence that CB1 agonists similarly induced PPI increases in rodents (Long et al., 2010; Stanley-Cary and Harris, 2002).

It is plausible that some of the psychoactive effects of cannabis in susceptible individuals are via effects on sensorimotor gating, both at early and later stages: in rats, agonists to the CB1 receptor, the principal target of THC in the brain, reduce prepulse inhibition, reversed by haloperidol (Schneider and Koch, 2002), and modulate emotional associative learning and memory formation (Laviolette and Grace, 2006). In humans Delta-9-tetrahydrocannabinol (THC), the key psychoactive constituent of cannabis, alters neural responses during basic sensory processing (Winton-Brown et al., 2011) and attentional oddball processing (Bhattacharyya et al., 2012), the latter particularly in the striatum and prefrontal cortex (Bhattacharyya et al., 2012). Alterations in so called ‘salience processing’ which may incorporate both early and late sensorimotor gating (Winton-Brown et al., 2014), induced by cannabis in susceptible individuals, may lead to the detection and prominence of stimuli that should have been filtered out, and induction of psychotic symptoms (Bhattacharyya et al., 2012; Kapur, 2003).

Cannabis did not appear to differentially alter sensorimotor gating at a later stage, as indexed by PPF. PPF was however significantly reduced in ARMS participants who were UDS-negative, suggesting that there may be later stage gating abnormalities in people at high risk of psychosis. These are similar to the findings from a study that found that PPI in the siblings of patients with schizophrenia was not different from controls, but that PPF was significantly reduced (Wynn et al., 2004).

We did not find relationships between either PPI or PPF and positive psychotic symptoms, consistent with previous literature suggesting that gating deficits relate more to cognitive deficits and disorganization (Braff et al., 1999; Karper et al., 1996). We did however find relationships of PPI and PPF sensorimotor gating with overall function, similar to Swerdlow et al. (2006) in participants with schizophrenia.

Our study had a number of important limitations, including a modest sample size; the subsample of UDS positive participants was particularly small with 5 and 6 in control and UHR groups respectively. The findings should be thus regarded as preliminary and require replication in larger samples. A limited sample size may also have limited the study's power to detect PPI deficits in the ARMS group regardless of cannabis use. However given the marked effects of cannabis use on PPI, in this sample future studies should include urinary drug it may be useful to collect more detailed information on the pattern and experience of substance use.