Borderline personality disorder (BPD) is a severe psychiatric condition characterized by a persistent pattern of emotional dysregulation, marked impulsivity, interpersonal disturbances, identity instability, and persistent non-suicidal self-injury (NSSI). This disorder has a 3 % prevalence in the general population (Lenzenweger et al, 2007; Grecucci et al, 2023) and is associated with significant impairment of patients’ psychological functioning (De Panfilis et al, 2019). Adolescence is considered a critical period of time for BPD (Ibraheim et al, 2017), as early treatment starting in adolescence may decrease symptom severity (A.M. Chanen et al, 2008). However, early diagnosis has been challenging due to the lack of understanding of its pathophysiology with no specific biomarkers to guide clinical decision-making.

The diagnosis of BPD is often a complex and difficult process due to several factors (Grecucci et al, 2022; Rao and Broadbear, 2019). One issue is that patient symptoms tend to fluctuate significantly, with different degrees of vulnerability (Choi-Kain et al, 2010; Zanarini et al, 2003). Another challenge is from the current classification system categorizing personality disorders in a categorical manner, whereas personality traits can be viewed on a spectrum in key domains (Langerbeck et al, 2023). Furthermore, there is considerable overlap in symptoms across different personality disorder categories (Morey et al, 2015). Lastly, diagnosing BPD depends on observable behavior, as no objective markers are available yet. An accurate early diagnosis is even more complex for adolescents with BPD (aBPD), as adolescent brain is still undergoing significant development and maturation.

Neuroimaging has emerged as a useful tool to assess structural and functional alterations of the brain (Bøen et al, 2014; de Araujo Filho et al, 2014; Bruehl et al, 2013), and has been recently applied to adults with BPD for the identification of possible biomarkers (Grecucci et al, 2022). The limited number of neuroimaging studies on aBPD have shown similar gray and white matter alterations as compared to adults (T. Takahashi et al, 2009; Jovev et al, 2008). These studies demonstrated multiple brain volume and surface alterations associated with clinical symptomatology in patients with aBPD (T. Takahashi et al, 2009; Jovev et al, 2008). Besides structural studies, brain functional MRI (fMRI) has also been used to characterize brain activity alterations in aBPD, which has identified three brain networks being affected in aBPD, i.e., the Central Executive network (CEN), the Default Mode network (DMN), and the Affective Network (AN) (Xiao et al, 2024). However, the neuroimaging studies conducted so far have several limitations, such as small sample size, unmatched gender ratio, and lack of control subjects (Lotfinia et al, 2020). Thus, the reported brain changes may not be representative of the patients with BPD in general.

This systematic review aimed to assess the current evidence on brain structural and functional alterations, and to examine whether these neuroimaging findings converge on common large-scale brain networks in aBPD. Through comprehensive literature review, we outlined three networks that were consistently implicated in aBPD, namely the CEN, the DMN, and the AN. We discussed how these network-level brain alterations may inform future hypothesis-driven research. In addition, we highlighted the potential value of emerging analytical approaches and artificial intelligence-based methods for identification of neuroimaging features associated with aBPD. While such methods have recently been used in adult BPD populations (Grecucci et al, 2023; Grecucci et al, 2022), they have not yet been systematically assessed in aBPD. Further research applying these techniques should improve our understanding of brain alterations associated with aBPD.