MSc Psychology Dissertation

An Exploration of the Relationship Between Adverse Childhood Experiences and Food Addiction Symptoms

Examining interoception and perceived stress as mediating pathways

Neringa Filipova

Institution

Anglia Ruskin University

Supervisor

Dr Jennifer Todd

Date

August 2026

Scroll to begin ↓

01 / Introduction

Anglia Ruskin University · Neringa Filipova

Foundations of the Research

Introduction & Background

Concept 01

Adverse Childhood Experiences

10-item measure established by Felitti et al. (1998)
Covers abuse (emotional, physical, sexual) and household dysfunction
4+ ACEs = 4–12× increased risk for alcoholism, depression, suicide

Concept 02

Food Addiction

Addictive-like response to highly processed foods (Brunault et al., 2020)
Measured using modified DSM-5 criteria (mYFAS-2.0)
Symptoms: loss of control, withdrawal, tolerance, cravings

The Problem

ACEs are linked to obesity and substance use — yet the mechanisms linking ACEs to food addiction remain unknown. Interoception may be a key pathway mediating this relationship.

02 / Theory

Anglia Ruskin University · Neringa Filipova

Theoretical Framework

Two complementary theories converge on interoception

01

Stress Sensitisation Theory

Repeated childhood adversity disrupts HPA axis regulation
Lowered stress thresholds → maladaptive coping
Ultra-processed foods become self-medication

02

Reward Deficiency Syndrome

ACEs cause dopaminergic dysregulation
Reduced reward from everyday activities
Seeking highly rewarding foods (sugar, fat, salt) for dopamine release

03

Interoception Integration

HPA dysfunction → difficulty perceiving hunger/satiety
Reward deficiency → heightened attention to external food cues
Vulnerability: eating for psychological rather than physical reasons

Both theories converge — HPA dysfunction impairs internal signal perception, while reward deficiency shifts attention externally, creating vulnerability to eating for psychological relief rather than physiological need.

03 / Aims

Anglia Ruskin University · Neringa Filipova

Aims & Hypotheses

Six hypotheses testing the ACEs → Food Addiction pathway

To examine the relationship between ACEs and food addiction symptoms in a UK adult sample — and test whether interoceptive accuracy, attention, and perceived stress mediate this relationship.

H1

Higher ACEs → higher food addiction symptoms

✓

H2

ACEs → ↓ interoceptive accuracy; ↑ interoceptive attention

H3

ACEs → ↑ perceived stress

H4

↓ accuracy, ↑ attention → ↑ food addiction symptoms

H5

↑ perceived stress → ↑ food addiction symptoms

H6

Parallel mediation model — interoception and stress mediate the ACEs–FA relationship

04 / Method

Anglia Ruskin University · Neringa Filipova

Methodology

Cross-sectional online survey with validated measures

Design & Participants

Design Cross-sectional, correlational online survey
Final N 67 (from 119 after exclusion)
Gender 89.6% female
Mean Age 25.73 years (SD = 9.53)
Population UK citizens/residents; no ED, diabetes, or appetite-affecting medication

Measures

ACE Questionnaire

Felitti et al., 1998

mYFAS-2.0

Brunault et al., 2020 — Food addiction

IAS

Mulder et al., 2026 — Interoceptive Accuracy

IAT

Gabriele et al., 2022 — Interoceptive Attention

PSS

Cohen et al., 1994 — Perceived Stress

Procedure

Platform Online via Qualtrics
Recruitment SONA system
Duration ~20 minutes completion time
Ethics Approved by ARU ethics committee; informed consent obtained

05 / Results

Anglia Ruskin University · Neringa Filipova

Descriptive Statistics

Sample characteristics and study variable distributions

Sample Characteristics

Characteristic	Value
Age (mean, SD)	25.73 (9.53)
Gender	89.6% Female
White British	41.8%
A-level education	56.7%

Study Variables

Variable	Mean (SD)	Range
ACEs total	2.24 (2.21)	0–8
mYFAS-2.0	29.54 (15.69)	13–78
IAS total	80.08 (12.40)	38–101
IAT total	48.11 (14.50)	21–80
PSS total	43.26 (7.45)	25–56

ACEs Distribution

<4 ACEs

73.1% (n=49)

4+ ACEs

26.9% (n=18)

06 / Results

Anglia Ruskin University · Neringa Filipova

Bivariate Correlations

Testing hypotheses 1, 2, and 4

Correlation Matrix (Pearson's r)

	ACE	MYFA	IAT	IAS	PSS
ACE	—	.416**	.111	-.221	.141
MYFA	.416**	—	.379*	-.068	-.129
IAT	.111	.379*	—	-.001	-.242
IAS	-.221	-.068	-.001	—	.084
PSS	.141	-.129	-.242	.084	—

* p < .05, ** p < .01 · Highlighted cells indicate statistical significance

✓

H1 Supported ACEs significantly correlated with FA symptoms (r = .416, p < .001) — moderate strength.

✗

H2 Not Supported ACEs not correlated with IAS (r = -.221, p = .088) or IAT (r = .111, p = .386).

◐

H4 Partially Supported IAT correlated with FA (r = .379, p = .002); IAS did not (r = -.068, p = .600).

07 / Results

Anglia Ruskin University · Neringa Filipova

ACEs Group Differences

Independent samples t-test on food addiction scores

<4 ACEs Group

25.96

SD = 13.42 · n = 49

4+ ACEs Group

38.58

SD = 18.32 · n = 18

Effect Size (Cohen's d)

0.78

Large effect · t(65) = 3.09, p = .003

Food Addiction Scores by ACE Group

<4 ACEs

M = 25.96

4+ ACEs

M = 38.58

Participants with 4+ ACEs had significantly higher food addiction symptoms. This supports the dose-response relationship — the more childhood adversity accumulated, the higher the food addiction symptoms in adulthood.

08 / Results

Anglia Ruskin University · Neringa Filipova

Mediation Analysis

Parallel mediation with 5,000 bootstrap resamples

Predictor

ACEs

IAS — Accuracy

B = -0.050, p = .777

IAT — Attention

B = 0.273, p = .341

PSS — Stress

B = -0.104, p = .478

Outcome

Food Addiction

Direct effect: B = 2.834**, p = .002

Pathway	Estimate	p-value	95% CI
Direct (ACEs → FA)	2.834	.002*	[0.883, 4.556]
Indirect via IAS	-0.050	.777	[-0.755, 0.210]
Indirect via IAT	0.273	.341	[-0.191, 1.404]
Indirect via PSS	-0.104	.478	[-1.022, 0.102]

H6 Rejected. While the direct effect of ACEs on food addiction remained significant (Total effect B = 2.953, p < .001), none of the indirect effects reached significance — all CIs included zero. The relationship operates through alternative mechanisms.

09 / Discussion

Anglia Ruskin University · Neringa Filipova

Discussion & Interpretation

Placing findings within the broader literature

Finding 01 · Supported

ACEs → Food Addiction

Consistent with Wiss et al. (2025): OR = 1.99 for 4+ ACEs. Supports stress sensitisation and reward deficiency theories — childhood trauma creates vulnerability to addictive eating.

Finding 02 · Not Supported

Interoception Not a Mediator

ACEs did not correlate with interoceptive dimensions. Contrasts with Schmitz et al. (2023). Possible explanation: self-report vs. behavioural measures may capture different constructs.

Finding 03 · Not Supported

Perceived Stress Not a Mediator

No significant correlation with ACEs or FA. PSS reliability issues (α = 0.575) may explain this. Sample characteristics (young, female, students) may also influence.

Implications

What This Means

Alternative mechanisms must be explored — emotion regulation, attachment, neurobiology
Trauma-informed approaches remain clinically indicated
Interoception training may still be beneficial despite non-mediation

10 / Reflection

Anglia Ruskin University · Neringa Filipova

Limitations & Future Research

Critical appraisal and directions for further inquiry

Limitations

Cross-sectional design

Cannot infer causality
Self-report measures

Recall and social desirability bias
Small sample (N = 67)

Reduced statistical power
Predominantly female (89.6%)

Limited generalisability
Convenience sample (students)

Not representative of general population

Future Directions

Longitudinal studies — establish temporal precedence and prospective prediction
Behavioural interoception measures — heartbeat detection tasks for objective data
Larger, diverse samples — community participants, males, older adults
Experimental manipulation — mindfulness training for eating behaviours
Qualitative research — lived experiences of ACEs and food relationships

11 / Conclusion

Anglia Ruskin University · Neringa Filipova

Conclusion

Key findings and implications for practice

✓

ACEs → Food Addiction

Moderate correlation (r = .416) · Large effect (d = 0.78) · Dose-response confirmed

✗

Interoception

Not a mediator · No ACEs correlation · No indirect effects via IAS or IAT

✗

Perceived Stress

Not a mediator · No correlation with ACEs or FA · Measurement issues likely

Clinical Implications

✓

Trauma-informed screening in clinical practice

✓

Nutrition counselling for individuals with ACEs

✓

ACE prevention as a public health priority

Childhood adversity becomes biologically embedded. Understanding how is critical for intervention.

Thank You

Questions & Discussion