go-v.orgdoi

The lasting impact of child abuse on neuroplasticity: an examination of perineuronal nets in the anterior cingulate cortex

Clémentine Hosdey · container not stated · 2026

  • DOI 10.82308/36978
  • Publisher OpenAlex
  • Type dissertation
  • Issued 2026

Not cached yet. This record was fetched live for this request and is stored after the lookup. Reopening the page should serve the cached copy.

· OpenAlex · dissertation · 2026

Abstract

Severe child abuse (CA) is a major environmental risk factor for depression and suicide. Depressive episodes in individuals with a history of CA are more likely to be chronic, recurrent, and treatment-resistant, and are accompanied by a two- to three- fold increased risk of suicide attempts. Despite decades of research, the cellular mechanisms through which early-life adversity alters developmental trajectories and contributes to such severe outcomes remain incompletely understood.

During childhood and adolescence, cortical development is characterized by windows of heightened plasticity, the closure of which coincide with the maturation of perineuronal nets (PNNs), specialized extracellular matrix structures that enwrap specific neurons. Converging evidence from animal models and human post-mortem studies suggests that PNNs are sensitive to early-life stress and may serve as a biological substrate through which CA becomes embedded in stress-sensitive brain regions. Previous work from our group has shown CA-associated increases in PNN density in the human ventromedial prefrontal cortex.

However, whether PNN organization within the anterior cingulate cortex (ACC), a region centrally involved in affect regulation and consistently implicated in depression and suicide, is impacted by a history of CA remains unknown.This thesis combines rigorous multi-rater annotation, deep learning–based image segmentation, and human post-mortem neuroanatomy to quantify PNNs in the ACC across well-characterized clinical groups. First, we assembled a multi-rater dataset of putative PNN annotations to assess inter-rater reliability.

We found that while detection agreement was fair (Fleiss’ κ = 0.35), reflecting variability in rater sensitivity, phenotype classification was robust (Krippendorff’s ⍺ = 0.92). Based on these metrics, we derived a high-confidence expert consensus ground truth to fine-tune and validate a tri-channel Cellpose-SAM model for automated PNN detection, which achieved performance matching or exceeding human pairwise consistency on held-out test tiles.In the biological application, the validated model was applied to post-mortem ACC sections from samples provided by the Douglas–Bell Canada Brain Bank (N = 44).

These samples were from psychiatrically healthy sudden-death controls (CTRL), individuals who died by suicide during a depressive episode without a history of CA (DS) and with documented severe CA (DS-CA). PNNs labeled with Wisteria floribunda lectin (WFL) and aggrecan (ACAN) immunostaining were quantified within manually delineated cortical layers. Total PNN density remained stable across groups, but DS-CA cases showed a selective increase in ACAN+ PNN density, significant in Layers II/III and V and with a trend toward higher density in Layer VI (p < 0.05, p ≈ 0.06), without a parallel change in WFL⁺ density.

Within-subject comparisons further revealed a layer-specific compositional shift in deep cortex (Layer VI), where the typical WFL-dominant profile seen in CTRL was narrowed in DS-CA. Together, these findings indicate that severe CA is associated with a compositional remodeling of the extracellular matrix in the human ACC, characterized by ACAN enrichment, and suggest that such alterations may contribute to persistent changes in plasticity-regulating microcircuits in depression and suicide

Open the resolved locator · 10.82308/36978

Metadata licence: OpenAlex metadata (CC0)

Registry
OpenAlex DOI index + API
Fetched
2026-11-20 03:27 UTC
Cache state
not cached yet (bulk-index+openalex)
Locator
https://hdl.handle.net/20.500.14905/138976
Metadata licence
OpenAlex metadata (CC0)
Scope
Metadata as returned by the resolver; abstracts are never generated here.