Chemosensory–Cognitive Profiles in Anorexia Nervosa

Technical Summary: Chemosensory–Cognitive Profiles in Anorexia Nervosa

Problem Statement
Anorexia nervosa (AN) is characterized by persistent energy restriction and pathological mental representations. While sensory processing alterations (chemosensory) and cognitive deficits (executive functions) are well-documented in AN, they have historically been investigated as isolated domains. The specific interaction between bottom-up sensory signals and top-down executive regulation remains poorly understood. It is unclear whether sensory changes in AN stem from primary peripheral deficits or are shaped by higher-order cognitive modulation. Furthermore, the heterogeneity of AN presentations—specifically the restrictive (ANR) and binge-eating/purging (ANB) subtypes, as well as the state of weight restoration (AN-REC)—suggests that a uniform model of impairment may be insufficient. This study aims to bridge this gap by investigating the direct relationships between chemosensory perception (olfaction and taste) and executive functioning across these distinct clinical profiles.

Methodology
The study employed a cross-sectional design involving 60 female participants divided into four groups:

• ANR (n=17): Restrictive-type AN (BMI ≤ 17.5 kg/m²).
• ANB (n=11): Binge-eating/purging-type AN (BMI ≤ 17.5 kg/m²), characterized by self-induced vomiting.
• AN-REC (n=10): Weight-restored AN (BMI 18.5–24.9 kg/m²) with a history of the disorder.
• NWS (n=22): Normal-weight subjects without a history of eating disorders.

Assessments:

1. Chemosensory Function:
   • Olfaction: Evaluated using the "Sniffin' Sticks" battery, measuring odor threshold (T), discrimination (D), and identification (I) to generate a global TDI score.
   • Gustation: Evaluated using the "Taste Strips" test, measuring sensitivity to sweet, sour, salty, bitter, and umami across four concentrations, yielding a total score (0–20).
2. Executive Functions: Assessed via the Test of Attentional Performance (TAP) battery and the Navon task, covering:
   • Inhibitory Control: Go/No-Go tasks (low and high demand).
   • Cognitive Flexibility: Set-shifting between verbal and non-verbal categories.
   • Working Memory: 2-back visual digit task.
   • Central Coherence: Navon task (global vs. local processing bias).
3. Psychometrics: Included BDI-II (depression), STAI (anxiety), and EDI-2 (interoceptive awareness).

Statistical Analysis:
Non-parametric tests (Kruskal–Wallis, Mann–Whitney U) were used for group comparisons due to violations of normality assumptions. Spearman's rank correlation coefficients were calculated to explore associations between sensory and executive variables within each group. Correlation matrices were visualized via heatmaps to identify patterns of sensory–cognitive coupling.

Key Results

• Group Performance Differences:

  • Gustation: The combined active AN group (AN-combined) showed significantly lower total taste scores compared to NWS, driven primarily by the ANB subtype. ANB patients had reduced sensitivity to salty, sweet, and bitter tastes. AN-REC patients also showed reduced scores for salty, bitter, and umami, while ANR patients showed no significant differences from controls.
  • Olfaction & Cognition: No significant group differences were found in overall olfactory performance or global executive function scores, with the exception of verbal flexibility differences between AN-REC and ANR.

• Sensory–Cognitive Correlation Profiles:

  • NWS (Controls): Exhibited a dense, homogeneous network of correlations. Taste performance was broadly associated with inhibitory control and central coherence. Sour taste sensitivity correlated with nearly all cognitive processes.
  • AN-Combined: Showed a more heterogeneous and selective pattern. Correlations were fewer but specifically tied to gustatory processing and executive functions (inhibition and flexibility). Notably, correlations with central coherence were largely absent.
  • Subtype Specificity:
    • ANR: Displayed a sparse correlation profile. Associations were limited to olfactory performance linked to working memory (reaction time and errors) and specific links between umami taste and verbal flexibility/local processing.
    • ANB: Exhibited the densest and most complex sensory–cognitive coupling. Higher taste scores were associated with enhanced flexibility and working memory but also with increased error rates in inhibitory and coherence tasks, suggesting a speed–accuracy trade-off. Olfactory identification also correlated with error rates in executive tasks.
    • AN-REC: Showed a marked reduction in correlation density compared to active AN. Significant associations were limited to salty taste sensitivity linked to cognitive flexibility and central coherence efficiency, and sweet taste linked to inhibitory control.

Key Contributions
The paper contributes a novel framework for understanding AN by shifting focus from isolated deficits to sensory–cognitive coupling profiles.

1. Top-Down Modulation: The findings suggest that chemosensory alterations in AN are not solely peripheral deficits but are significantly influenced by top-down cognitive processes. The altered correlation structures in AN imply that executive systems actively shape sensory processing.
2. Subtype Differentiation: The study provides empirical evidence that ANR and ANB represent distinct neurocognitive configurations. ANR appears characterized by stable, automated top-down control (decoupled from acute sensory input), whereas ANB is characterized by variable, sensory-driven cognitive control with a speed–accuracy trade-off.
3. Persistence of Alterations: The AN-REC group demonstrated that while some associations normalize, specific couplings (e.g., salty taste and executive efficiency) persist after weight restoration, suggesting a reconfigured, compensatory architecture rather than a full return to baseline.

Significance and Claims
The authors claim that their data support an integrated view of sensory and executive functioning in AN. They posit that the disorder is best conceptualized by subtype-specific configurations of neurocognitive integration rather than uniform impairments.

• Clinical Translation: The identification of these distinct profiles suggests that interventions should be personalized. For instance, individuals with the binge–eating/purging phenotype, who show severe gustatory compromise and high sensory–cognitive coupling, might benefit from "bottom-up" sensory rehabilitation (e.g., flavor-enhanced foods) to stabilize intake, rather than relying solely on executive control strategies.
• Limitations: The authors modestly note that their correlation analyses are exploratory and descriptive. They acknowledge potential confounds such as psychotropic medication use and the lack of food-specific stimuli in cognitive tasks. They emphasize that the results generate hypotheses regarding the mechanisms of sensory–cognitive integration rather than confirming definitive causal pathways.