Being born preterm with very low birthweight (VLBW, birthweight ≤1,500 g) or small for gestational age (SGA, birthweight <10th percentile) at term constitute important risk factors for later cognitive difficulties (Aarnoudse-Moens et al., 2009).

For individuals born with VLBW and/ or very preterm (born <32 weeks of gestation), there is evidence for impaired cognitive performance including lowered IQ scores, executive, and attentional difficulties in childhood (Mu et al., 2008; Aarnoudse-Moens et al., 2009; Murray et al., 2014; Breeman et al., 2015; Farajdokht et al., 2017; Córcoles-Parada et al., 2019), adolescence (Brydges et al., 2018; Twilhaar et al., 2020), and adulthood (Nosarti et al., 2007; Pyhälä et al., 2011; Eryigit Madzwamuse et al., 2015; Kroll et al., 2017; Eves et al., 2020; Perrone et al., 2020; Eves et al., 2021), and a dose–response-relationship between birthweight and IQ has been indicated (Gu et al., 2017).

Whether the described cognitive difficulties persist into adulthood has been a question of some controversy, and while an outgrowing of sequelae due to brain plasticity has been suggested (Curtis et al., 2002; Schneider et al., 2014), most researchers find support for lasting cognitive difficulties in individuals born preterm with VLBW (Pyhälä et al., 2011; Heinonen et al., 2013; Eryigit Madzwamuse et al., 2015; Kroll et al., 2017; Linsell et al., 2018; Stålnacke et al., 2019).

Being born SGA at term may indicate intrauterine growth restriction, which has been associated with later neurocognitive and neurodevelopmental abnormalities (Sharma et al., 2016; Sacchi et al., 2020) and poorer school performance (Lindström et al., 2017). In a review, de Bie et al. (2010) found that children and adolescents born SGA at term had lower IQ scores and more cognitive difficulties than controls. Likewise, Yi et al. (2016) reported lower verbal IQ scores in a group of term-born SGA children at 8–16 years of age. Difficulties on tests of attention and executive function in SGA individuals at 20 years of age was observed in a study by Suffren et al. (2017).

Poor cognitive function is found to be associated with lower real-life achievements in individuals born very preterm or with VLBW (Kroll et al., 2017). Furthermore, an increased risk of social difficulties and poor mental health is demonstrated for VLBW individuals (Kroll et al., 2018; Kajantie et al., 2019; Anderson et al., 2021; Ni et al., 2021). For extremely low birth weight (ELBW) individuals in middle childhood, frequently reported diagnoses include attention deficit/hyperactivity disorder (ADHD), autism spectrum disorders (ASD) and anxiety disorders (Johnson and Marlow, 2011; Anderson et al., 2021). Increased risk for mental health problems in adults born with low birthweight or born preterm was also found in a large meta-analysis by Pyhälä et al. (2017) as well as Anderson et al. (2021) and in register-based population studies (Nosarti et al., 2012; Class et al., 2014; Persson et al., 2020; Robinson et al., 2022).

In a previous follow-up of the present cohort in early adulthood (at 19–20 years of age), the preterm VLBW group had lower IQ scores than the control group (Løhaugen et al., 2010) and displayed problems related to memory (Aanes et al., 2015), attention and executive functions (Sølsnes et al., 2014; Østgård et al., 2016). The term born SGA group had lower IQ scores (Løhaugen et al., 2013) and modestly poorer scores than the control group on attention and executive functions, with more significant differences in the memory domains (Østgård et al., 2014). We found increased frequency of mental health problems at 14 and 20 years of age in the VLBW group, specifically anxiety disorders, attention problems and autistic traits (Indredavik et al., 2004; Lund et al., 2011), and furthermore increased frequency of psychiatric disorders and symptoms in both the preterm VLBW and the term born SGA group at 26 years of age (Lærum et al., 2017, 2019).

These reports raise the question whether lower cognitive reserve is associated with adult mental disorders in preterm VLBW, and term born SGA adults.

In this study, we aimed to assess neurocognitive function as well as its associations with the typical profile of psychiatric symptoms and diagnoses in adults born preterm at VLBW or born term SGA compared to controls, specifically anxiety/internalizing problems, inattention, and autistic traits. We expected adults born preterm with VLBW or term SGA to perform poorer than controls on tests of neurocognitive functions. Furthermore, we anticipated psychiatric symptoms/diagnoses to be associated with lower cognitive functioning.

In accordance with our expectations, adults born preterm with VLBW or SGA at term had several cognitive difficulties compared with controls. The VLBW group had reduced Full scale, Verbal and Performance IQ, lowered performance on aspects of visual learning, lowered psychomotor speed, poorer performance on tests measuring attention switching, verbal fluency, spatial working memory, visual processing, social cognition, and fine motor function. The SGA group had reduced Performance IQ and psychomotor speed. Associations between neurocognitive function and mental health were present in the VLBW group, where spatial working memory was associated with the more frequent presence of anxiety disorders, and with higher levels of internalizing problems, attention problems and autistic traits. Furthermore, a lower Full scale IQ was associated with more attention problems when adjusting for sex and parental SES (available case analysis). For the SGA group and the control group, no associations were found between cognitive function and mental health.

The results on cognitive function in this study are in accordance with previous studies on adults born preterm at VLBW, which reported difficulties related to IQ, executive function, attention and visual memory (Pyhälä et al., 2011; Eryigit Madzwamuse et al., 2015), and support the notion that cognitive difficulties persist into adulthood, as also shown for IQ in a recent individual participant data meta-analysis (Eves et al., 2021). Weak spatial working memory was furthermore seen in a study by Suikkanen et al. (2021) on the Cogstate Groton Maze Learning Test, and the authors argue that these difficulties could be related to the smaller size of the hippocampi, which was also found in the present sample at age 19–20 (Aanes et al., 2015). Reduced hippocampi volume was, moreover, found for the present preterm born sample at 26 years of age, but without significant correlation between spatial working memory and reduced hippocampal subfield volumes (Aanes et al., 2020).

Furthermore, the present study extends the current knowledge base on cognitive function in individuals with VLBW by revealing difficulties in visual learning and memory, verbal fluency, and social cognition. Visual memory difficulties in the VLBW group might be related to poor working memory function and thus difficulties in the initial encoding of the material. Weak verbal fluency, although mainly indicating difficulties in executive function such as initiation and generation, might suggest verbal challenges in adult age.

Of importance is the finding that individuals born SGA at term showed some cognitive difficulties in adult age with lower Performance IQ scores and reduced psychomotor speed compared to controls. This extends the findings from Viggedal et al. (2004) who found a slight reduction in IQ scores in young adulthood in individuals born SGA at term compared to controls, and underlines the importance of devoting attention to cognitive function in individuals born SGA in adult age.

Social cognition in adults born preterm with VLBW or SGA at term has not previously been investigated using the CANTAB. As described, the VLBW group performed below the controls on this task and are thus less able to differentiate between rapidly presented emotional facial expressions. This finding fits within the previous reports of poorer social skills in adults born with VLBW (Montagna and Nosarti, 2016). However, as the images used in the emotion recognition task are presented very briefly, these results could in part be due to problems with attention shifting and processing speed.

In the VLBW group, the associations between spatial working memory difficulties and symptoms belonging to the typical mental health profile consisting of elevated levels of attentional difficulties, symptoms of anxiety and poor social function (Johnson and Marlow, 2011) are especially noteworthy. Working memory is presumed to be important for everyday performance and a stronger predictor of academic achievement than IQ (Alloway and Alloway, 2010). Difficulties in spatial working memory was observed in very- and moderately preterm children at the age of 11 (Fitzpatrick et al., 2016), and in a NICU sample (of whom 78% were born preterm) measured with the CANTAB at age 9–14 (Curtis et al., 2002). Arthursson et al. (2017) found atypical neuronal activation in a group of very preterm adolescents during spatial working memory tasks, and the authors proposed a delayed maturation of the dorsolateral prefrontal cortex in the preterm group. Spatial working memory has also been found to be reduced in children with ASD (Zhang et al., 2020), and spatial working memory deficits are suggested to bring about social difficulties through problems in locating peers in context (Jiang et al., 2014). It is not unlikely that this might also affect social cognition in persons born at VLBW.

Furthermore, estimated IQ was associated with attention problems in the VLBW group after adjusting for sex and Parental SES. This is in accordance with findings from the general population and has also been suggested for extremely low birthweight individuals (Mathewson et al., 2017). The impact of IQ on psychopathology extends findings from a previous study on this VLBW cohort, where those exhibiting a psychiatric disorder at 15 and 19 years, or who developed psychopathology between these ages, had significantly lower IQ scores than VLBW participants without a psychiatric disorder at age 19 (Botellero et al., 2017). In the SGA group and the control group, however, there were no associations between cognitive function and mental health, despite increased prevalence of psychopathology in the SGA group at 26 years (Lærum et al., 2017). We need to bear in mind that the SGA group is heterogeneous, as some of them may be constitutionally small but otherwise healthy, and some may be growth restricted.

Taken together, our results regarding spatial working memory indicate a possible relationship between this function and anxiety, attention deficit and autistic traits in the VLBW group. The contribution of cognitive difficulties in general in a range of psychiatric disorders is well known (Millan et al., 2012), and individuals born preterm with VLBW may thus constitute a high-risk group for developing psychopathology. We speculate that cognitive difficulties affecting spatial working memory may imply reduced capacity to handle information and solve problems as well as reduced capacity to learn mental and social skills. Furthermore, altered brain development from pre- or perinatal exposures in VLBW individuals might result in vulnerability to both cognitive difficulties and psychopathology, though emerging or manifesting clinically at different ages. As cerebral MRI alterations are demonstrated in VLBW individuals (Volpe, 2009; Li et al., 2015), we speculate that SWM is specifically vulnerable to less optimal neurodevelopmental conditions in these individuals, and consequently, associations with mental health measures were found only in the VLBW group. Moreover, low SES and adverse environmental factors are found to be associated with being born with low birthweight (Blumenshine et al., 2010), as well as both cognitive difficulties and psychopathology (Gur et al., 2019). Overall, early detection of cognitive difficulties in help-seeking individuals born with low birthweight and concurrent psychopathology is important, as these individuals might require tailored therapeutic interventions targeting both functions.

We must bear in mind that the obstetric and neonatal care has improved considerably since the pre-surfactant period in the 1980s, when this VLBW cohort was born. Improved medical care, population health and socioeconomy have led to increased survival rates (Howson et al., 2013). Despite these advances, a meta-analysis displayed that extremely or very preterm children born in the antenatal corticosteroid and surfactant era show large deficits in intelligence, and no improvement in cognitive outcome was observed between 1990 and 2008 (Twilhaar et al., 2018). Another meta-analysis showed that children (4 to 17 years) born very preterm had medium to large deficits in three major cognitive domains: executive function, processing speed, and intelligence, with increasing severity with decreasing gestational age and/or birthweight (Brydges et al., 2018). Thus, even though our results may not be generalized to comparable gestational ages or birthweights in more recent cohorts, the results may still be relevant for those born with extremely low birthweight at the border of viability.

The strength of this study is the prospective design investigating both a preterm VLBW group and a term SGA group compared to a control group in adult age assessed with a comprehensive battery of neurocognitive tests. The mental health assessment was performed with validated instruments, and for the semi structured interview, means to ensure interrater reliability were applied. Apart from parental SES, our information is limited on environmental factors that may contribute to both cognitive development and mental health. However, all participants have grown up in Norway with equal access to health care and schooling. Furthermore, the fact that we had measures of SES for only a reduced sample, limits the applicability of this adjustment. Nonetheless, we chose to adjust for SES when analyzing the associations between neurocognitive function and mental health, since mental health is found to be strongly associated with SES (McLaughlin et al., 2011). Moreover, since the participants exhibited cognitive difficulties, self-report might not be the most reliable method for detecting psychiatric disorders/symptoms. The use of an abbreviated intelligence test (WASI) is a further limitation. Compared to a follow-up of the same sample at age 19–20 years (Løhaugen et al., 2010, 2013), when the Wechsler Adult Intelligence Scale, 3rd Edition (WAIS-III) (Wechsler, 1997a) was used, the IQ scores increased substantially and beyond what we would have expected in all three groups. Since there do not seem to be a similar increase in test scores from 19 to 20 years of age to 26 years of age on other neuropsychological tests such as the WMS-III (Østgård et al., 2014; Aanes et al., 2015), this discrepancy likely indicates a difference in the two IQ tests rather than an actual increase in IQ, and this elevated IQ-score might have masked associations between cognitive impairments and mental health problems. The use of the CANTAB set in “clinical mode” might constitute another weakness of this study, since a ceiling effect might have occurred. Since we used data from a large number of tests, we chose to reduce the risk of making a Type 1 error by applying a p-value of.01 as significance level. The use of this strict significance level might, however, have increased our risk of making a Type 2 error. The relative low sample size might furthermore involve increased risk of making Type 2 errors.

In the present study, a broad neuropsychological test battery was used to examine the neurocognitive functions of adults born preterm with VLBW or term SGA. The VLBW group performed significantly poorer than the control group on tests measuring several cognitive domains, while the SGA group performed at a level between the VLBW group and the control group yet indicating that they may have mild and unrecognized cognitive challenges. For the VLBW group, the association between spatial working memory and difficulties with attention, anxiety, and social function should be further explored in research and taken into consideration in the treatment of such psychiatric problems in this group.

The datasets generated and/or analysed during the current study are not publicly available as permission has not been provided by participants nor the Ethical Committee. Requests to access the datasets can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by the Regional Committee for Medical and Health Ethics (REK Central Norway). The participants provided their written informed consent to participate in this study.

SW contributed to the analyses and interpretation of data, drafted the initial manuscript, revised, and approved the final manuscript as submitted. AL carried out the analyses and interpretation of data, drafted the initial manuscript, revised, and approved the final manuscript as submitted. SR took part in and supervised the psychiatric assessment, reviewed, and revised the manuscript and approved the final manuscript as submitted. KE participated in conceptualization and designing the study, contributed to the analyses, reviewed, and revised the manuscript and approved the final manuscript as submitted. SL supervised statistical analyses, reviewed, and revised the manuscript and approved the final manuscript as submitted. A-MB participated in conceptualization and designing the study, reviewed, and revised the manuscript and approved the final manuscript as submitted. JS participated in conceptualization and designing the study, reviewed, and revised the manuscript and approved the final manuscript as submitted. MI participated in conceptualization and designing the study, coordinated and supervised data collection, supervised the analyses and interpretation of data, and revised and approved the final manuscript as submitted. All authors contributed to the article and approved the submitted version.

The study was supported by the Norwegian Research Council and the Joint Research Committee between St. Olavs Hospital and Faculty of Medicine and Health Sciences, NTNU. AMWL was funded by the Liaison Committee for education, research and innovation in Central Norway.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.