Understanding Mental Health Through Fully Idiographic Networks

Andreoli, G., Rafanelli, C., Hofmann, S.G. et al. A Systematic Scoping Review of Fully Idiographic Network Analysis in Mental Health. Cogn Ther Res (2025). https://doi.org/10.1007/s10608-025-10674-2

Introduction

Mental health isn’t a one-size-fits-all equation. We all experience emotions, thoughts, and behaviors uniquely, yet much of traditional psychological research often relies on “nomothetic” or group-level data. This approach, while valuable, can miss the intricate, personalized dynamics of an individual’s mental health. Enter Fully Idiographic Network Analysis (FINA) – a brilliant method that focuses on understanding mental health at the deeply personal, “N=1” level.

A recent scoping review by Andreoli et al. (2025) offers the first comprehensive look at how FINA is being used in mental health research, highlighting its promising potential while also pointing out critical areas for improvement. This deep dive into 43 studies (N=43) published between 2011 and 2025 reveals a vibrant, rapidly evolving field grappling with both innovation and inconsistency.

What is FINA and Why Does it Matter?

Imagine your symptoms – anxiety, low mood, sleep problems – not as isolated issues, but as interconnected nodes in a complex web. FINA helps map these individual symptom networks by collecting intensive longitudinal data from a single person over time. This personalized map can then reveal which symptoms trigger others (a temporal network) or which tend to co-occur intensely (a contemporaneous network). The ultimate goal is to identify unique “drivers” or “hubs” within an individual’s network, offering personalized targets for treatment that couldn’t be seen in group-level data

For instance, in a contemporaneous network, strength centrality (the sum of absolute edge weights connected to a node, indicating how intensely a symptom co-occurs with others) could highlight that for one person, somatic arousal and panic anticipation are tightly linked. In a temporal network, instrength (sum of incoming absolute edge weights, showing how much a symptom is predicted by others) and outstrength (sum of outgoing absolute edge weights, showing how much a symptom predicts others) might reveal that low mood consistently predicts anxiety, suggesting mood as a potential intervention point.

The Landscape of FINA Research: A Snapshot

The review found that FINA research is characterized by considerable heterogeneity across almost all aspects. Here’s what stood out:

• Intensive Data Collection Dominates

  • The majority of studies, 65.1% (n=28), used Ecological Momentary Assessment (EMA) or Experience Sampling Methods (ESM) – where participants report on their experiences multiple times a day, often via electronic devices (96.4%, n=27 of these studies). This intensive approach is crucial for capturing dynamic symptom interplay, with the average completion rate for assessments, when reported, being 84.3%.

• Many Participants, Few True N=1s

  • Despite the “idiographic” (individual-focused) label, only 16.3% (n=7) of studies were true single-subject designs. The mean sample size per FINA analysis was 41.5 (SD=63.8), but surprisingly, FINA results were only actually reported for a maximum of 133 participants (Mean=18.2, SD=33.8) across all studies. This suggests that many studies applying FINA still aggregate data or only present individual results illustratively, potentially limiting deep personalized insights.

• R is King

  • When it comes to analysis, R (93%, n=40 studies) is the overwhelming choice of software, utilizing packages like graphicalVAR and psychonetrics.

• GVAR is the Go-To Model

  • The Graphical Vector Auto-Regressive (GVAR) model (a time-series model capturing both direct and instantaneous effects between variables) was the most frequently employed FINA model, used in 34.9% (n=15) of studies. Other models like Dynamic Time-Warp (DTW) (16.3%, n=7) and Structural Equation Modeling (SEM) variants (14% total, n=5) also emerged as flexible alternatives, especially when data assumptions for GVAR (like stationarity) are violated.

• Assumptions Often Overlooked

  • A significant concern is the infrequent testing of key statistical assumptions. For example, normality (data distributed in a bell-shaped curve) was tested in only 8.8% (n=3) of studies assuming it, and stationarity (statistical properties not changing over time) in only 25% (n=8). This raises questions about the validity of some reported findings. Similarly, topological overlap (when nodes measure overlapping constructs) was assessed in a mere 16.2% (n=6) of applicable studies.

• Stability is Rarely Checked

  • Assessing network stability (robustness to sampling error), usually via bootstrap methods (resampling to estimate statistic distribution), was performed in a meagre 11.6% (n=5) of studies. This is crucial to ensure that the identified networks aren’t just artifacts of a particular data collection period.

• Open Science Practices Are Emerging but Inconsistent

  • While analysis code was shared in 58.1% (n=25) of studies, data sharing was lower (30.2%, n=13), likely due to privacy concerns with sensitive individual-level data. Preregistration (publicly registering study plans before data collection), a cornerstone of transparency, was exceptionally rare, reported in only 7% (n=3) of studies.

Recommendations for More Rigorous FINA

The review authors provide a detailed checklist (Table 2) and recommendations to steer FINA toward greater scientific rigor and clinical utility.

Key suggestions include:

1. Standardize Data Collection

   For reliable edge detection, target at least 75 timepoints for networks with up to 6 nodes. Fixed-interval EMA/ESM is often preferred for models assuming equidistant data points.

2. Test Assumptions Religiously

   Routinely check for normality and stationarity. If assumptions are violated, consider models like DTW or Contingency measure-based Network (ConNEct) that are more robust to such issues. Address overnight lag (the time gap between the last assessment of one day and the first of the next) to maintain time-series integrity.

3. Manage Network Complexity

   Limit the number of nodes (individual symptoms or variables) in a network, with a mean maximum of 12.3 nodes observed in current research, to avoid overfitting (a model fitting too closely to training data, losing generalizability) and enhance interpretability.

4. Prioritize Stability and Comparison

   Implement data-dropping bootstrap or similar methods to assess network stability. Move beyond purely visual comparisons to formal statistical approaches like the Individual Network Invariance Test for comparing networks across or within individuals.

5. Embrace Transparency

   Improve reporting practices by providing detailed participant characteristics and individual-level FINA results. Crucially, promote open science by preregistering studies and sharing de-identified data and analysis code, while maintaining patient privacy.

What This Means for Mental Health Professionals

For clinicians, FINA holds immense promise for personalized care. Understanding a patient’s unique symptom network could lead to highly targeted interventions. However, the review underscores that FINA is still an emerging science.

• Hypothesis Generation

  View FINA-derived insights, like a high strength centrality for a particular symptom, as powerful hypotheses for personalized treatment, rather than definitive causal claims.

• Informed Assessment

  If considering FINA, advocate for robust data collection protocols, such as fixed-interval EMA/ESM with enough data points to reliably build networks.

• Mindful Interpretation

  When interpreting network diagrams, discuss with patients how their individual symptoms interact and how changes in centrality post-treatment might indicate progress.

• Ethical Vigilance

  Ensure stringent data privacy and de-identification, and obtain explicit consent for the intensive data collection required by FINA.

The journey to fully personalized mental health care is complex, but FINA offers a powerful compass. By addressing current methodological gaps and embracing rigorous, transparent practices, researchers can unlock FINA’s full potential to tailor interventions and truly understand the individual labyrinth of mental health.