This study is a seven-year follow-up study of 57 patients with first-episode schizophrenia in two neighboring Norwegian counties. For a description of the baseline study, including inclusion and exclusion criteria, see Haug et al. (2012). In the follow-up study, 35 patients (61%) agreed to participate. For a description of the follow-up study, see Svendsen et al. (2018). All participants provided informed consent to participate at both time points. The study was approved by the Regional Committee for Medical Research Ethics and the Norwegian Data Inspectorate (Figure 1).

Diagnoses were set by using the Structured Clinical Interview for DSM-IV Axis I disorders (SCID module I, chapter A–E; First et al., 1996), with additional information from medical charts.

The interview for the Positive and Negative Syndrome Scale (SCI-PANSS) (Kay et al., 1987) was used to measure psychotic symptoms. The use of antipsychotics at baseline and follow-up was measured as Defined Daily Dosages (DDD; WHO, 2023).

Clinical psychologists conducted the baseline assessments, and two trained research assistants performed the follow-up assessments. An experienced neuropsychologist and researcher (MØ) trained and supervised them. All subjects were examined separately, and the neurocognitive tests were given in the same order. The patients were tested when they were clinically stabilized and not in an acute episode of illness. At baseline, a more comprehensive test battery was used. The total time for all assessments was approximately 3 h at baseline and 2 h at follow-up. The current estimated IQ at baseline was assessed with the Wechsler Abbreviated Scale of Intelligence (Wechsler, 2007). The remaining tests cover domains that are sensitive to neurocognitive dysfunction in psychosis (Green et al., 2004):

Psychomotor speed: Digit Symbol from Wechsler Adult Intelligence Scale (WAIS-III; Wechsler, 1997; Wechsler, 2003). For 120 s, the participant has to fill in blank spaces with the symbol corresponding to the number above the blank space as rapidly as possible. The number of squares correctly filled in determines the score.

Immediate verbal memory: Part 1 of the Logical Memory Test from the Wechsler Memory Scale [WMS] III (Wechsler et al., 2008). In this test, the subjects are required to immediately recall details of two short verbal stories.

Executive functions: Letter Number Span from WAIS-III (Wechsler, 1997, 2003). The task involves hearing a series of letters and digits and then reporting back the stimuli in alphabetical order and ascending numerical order. The result is a total of correctly recalled trials. This task requires auditory working memory.

We included selected measures from the Delis-Kaplan Executive Function System (D-KEFS; Delis et al., 2001). In the Colour-Word Interference Test (CWIT), Condition 3, the participant has to inhibit an overlearned verbal response when naming dissonant ink colors in which the words are printed. In CWIT Condition 4, the person has to switch back and forth between naming the dissonant ink colors and reading the words. Condition 3 measures inhibition, and condition 4 measures cognitive flexibility. Completion time in seconds was examined.

For all tests, standard scores or T-scores according to norms were used (See Table 1).

SDs were assessed with the Examination of Anomalous Self-Experience (EASE) (Parnas et al., 2005), a standard measure focusing specifically on SDs, comprising five domains: 1. Cognition and stream of consciousness. 2. Self-awareness and presence. 3. Bodily experiences. 4. Demarcation/transitivism. 5. Existential reorientation. For a detailed description of this assessment, see Svendsen et al. (2018).

The EASE manual usually aims to capture the lifetime experiences of SDs, and at baseline, we registered only lifetime experiences. Since we aimed to measure change and stability in SDs at follow-up, we rated SD experiences during the last 2 years separately from lifetime experiences at follow-up. We used the information about the last 2 years in the current analyses. EASE change was calculated as EASE at baseline minus EASE at follow-up.

All analyses were performed with the statistical package SPSS, version 28.0 (SPSS Inc., Chicago, IL, United States). Mean and standard deviations are reported for continuous variables and percentages for categorical variables. All variables of interest were examined for deviations from normality. Neurocognitive indices and EASE scores at both time points were normally distributed. Antipsychotic use at both time points were skewed, and were transformed to their natural logarithm before they were entered into parametric analyses. Partial correlation analyses were used to investigate possible associations between neurocognitive measures at baseline and EASE change and EASE total score at follow-up, corrected for the use of antipsychotics. Paired sample statistics were used to measure differences in neurocognitive function and SDs from baseline to follow-up.

Our main findings were that poorer baseline executive functions (inhibition and cognitive flexibility) in patients with schizophrenia were associated with higher levels of SDs at 7 years of follow-up. Further, at baseline, better executive functions, particularly cognitive flexibility and working memory were associated with a more significant decrease in SDs during the follow-up period. We also found that immediate verbal memory and psychomotor speed were not associated with changes in SDs or SDs at follow-up. In the previous cross-sectional baseline study with more patients included (N = 57), however, we found that high levels of SDs were significantly associated with poorer verbal immediate memory (Haug et al., 2012). We hypothesized that deficiencies in immediate verbal memory could lead to problems understanding, directing, remembering, and reasoning about one’s thoughts and self-knowledge, functions linked to various aspects of SDs. However, our present follow-up results indicate that immediate memory difficulties are less critical for how SDs develop over time. Instead, we found that executive dysfunctions at baseline were associated with higher and more stable SDs over time (i.e., less reduction in SDs over time and higher SDs scores at follow-up). Short-term (immediate) memory and working memory have been used to refer to the maintenance and the maintenance plus manipulation of information, respectively. However, correlational studies have been unable to distinguish the two constructs consistently, and there is evidence of a significant, if not complete, overlap (Aben et al., 2012). Thus, working memory is not entirely distinct from short-term or immediate memory as it includes short-term memory (Cowan, 2008).

We expected that SDs should have some form of neurocognitive correlates, but we found very few significant correlations. However, phenomenological psychopathology is a purely subjective, experiential scientific domain of mental life, arising exclusively from patients’ self-reports, and only possible to assess through in-depth interviews. The neurocognitive fields of psychopathology are, on the contrary, test-based, objectively assessed sets of mental functions. This fact may leave its imprint on potential relationships between the two scientific domains.

Executive functions generally allow us to anticipate outcomes and adapt to changing life situations and are essentially the conscious regulation of thought, emotion, and behavior (Diamond, 2013). These functions help us focus our attention, resist distraction, preserve flexible thinking, set-shifting, and keep a thought or an image in mind (Prebble et al., 2013; Alderson-Day and Fernyhough, 2015). They provide the capacity to bind experiences across time, to perceive the present moment as both a continuation of our past and as a prelude to our future (Barkley, 2004; Prebble et al., 2013). Thus, executive functions sub-serve aspects of self-regulation, and impairments in executive functions may be associated with impairing aspects of the self’s ability to function and maintain a sense of constancy and continuity. It is reasonable to believe that the executive sub-functions impact “continuity” in the sense of selfhood, emotional self-regulation, and inner speech (Prebble et al., 2013; Alderson-Day and Fernyhough, 2015).

The executive function tasks used in the interference-test in this study require a weighting in favor of “top down” control. Hierarchical predictive processing has recently gained substantial attention as a computational neuroscientific approach to psychotic symptoms (Fletcher and Frith, 2009; Seth, 2013; Powers et al., 2017) and has also been applied to SDs (Seth et al., 2011; Clowes and Gärtner, 2018; Sass et al., 2018). On this account, top-down signals convey predictions based on prior expectations, while bottom-up signals convey prediction errors. The brain uses previously generated high-level causal models to attempt to predict sensory activity continually. The prediction error is the discrepancy between the reward and its prediction. The precision-weighting of predictions and prediction error is mediated by noradrenergic, cholinergic, glutamatergic, and gamma-aminobutyric acidergic function (Powers et al., 2016). Difficulties with the balance and weighting of top-down and bottom-up processing may have a role in the onset and maintenance of psychotic symptoms and SDs, according to new models (Clowes, 2018; Clowes and Gärtner, 2018; Nelson et al., 2019). The current findings, which are based on the relationship between EASE scores at follow-up and the interference test, provides the first direct empirical data that predictive coding (specifically, disturbance in the “top-down,” prior-based interpretation of information) may play a role in SDs, a finding worthy of further investigation.

Executive dysfunctions were not related to SDs at baseline in our patients. However, poorer executive functions at baseline were associated with more SDs at follow-up. Because these functions are essential for self-regulation, extended and persistent impairments in executive functions in everyday life may reflect and influence the development of SDs over time. Although we found associations between executive functions at baseline and SDs at follow-up, there may be mediating or moderating factors that we did not investigate. We do not know any previous studies investigating the longitudinal relationship between neurocognitive impairments and SDs, and further studies are necessary to generate more knowledge regarding this topic. The current study sample is small, with the possibility of finding clearer associations between neurocognitive impairments and SDs in the context of larger samples. Future work should also look at a broader range of neuropsychological functions such as aberrant salience, time awareness, source monitoring, hierarchical predictive processing, and multisensory integration – i.e., go beyond the “standard” neuropsychological battery (Nelson et al., 2019). Future findings across the phenomenological and neurocognitive domains could represent a significant advance in understanding schizophrenia and potentially improve early identification and intervention strategies.

In conclusion, executive functions may influence the development of SDs, the core phenomenological features of schizophrenia, over time, and possibly the other way around.