Herman 2026 — Reduced differentiation between physiological and emotional states in FM (emBODY)

One-paragraph summary

The empirical FM-direct anchor for the interoception/predictive-coding mechanism. Two independent cross-sectional studies of FM patients vs. age- and gender-matched pain-free controls. Participants completed the emBODY task: a body-mapping protocol in which they color-code the topographic distribution of bodily sensations associated with different internal states (specific emotions, neutral state, pain-related states, physiology-related states, fatigue). The central result, replicated across both studies: linear discriminant analysis showed lower classification accuracy of body-sensation maps in the FM group than in controls — i.e., FM patients' body maps for different emotional and physiological states are less differentiated from each other. Concurrent self-report measures showed FM patients had higher alexithymia, higher awareness of bodily signals, more negative interpretation of ambiguous bodily sensations, and greater self-reported interoceptive difficulties. The combined picture is amplified perception of bodily signals coupled with poorer differentiation of states — direct empirical instantiation of the interoceptive-inference deficit predicted by the Strube 2026 framework.

Claims as triples

• interoceptive_inference — modulates → central_sensitization [evidence: lower LDA classification accuracy in FM body-sensation maps; confidence: emerging]
• alexithymia — present_in → fm_central_only [evidence: higher alexithymia in FM cohort; confidence: emerging]
• interoceptive_inference — modulates → widespread_pain [evidence: misattribution of non-noxious and emotional signals as pain; confidence: emerging]
• predictive_coding_failure — present_in → fm_central_only [strengthened: empirical signature of poor state differentiation matches Strube's predicted phenotype; confidence: emerging]

Methods note

Two independent cross-sectional studies, both with FM patients vs. age/gender-matched pain-free controls. emBODY task (Nummenmaa-style body-sensation mapping). Self-report measures: alexithymia (TAS-20 likely), bodily sensation interpretation, interoception (MAIA likely). Linear discriminant analysis tested whether different emotional/physiological states produced statistically distinct body sensation patterns. Authors: Herman (likely Polish/Australian collaboration based on co-author names — Szczawińska, Berryman/Stanton at U South Australia).

Limitations

• Preprint, not peer-reviewed. Replication exists within the paper (two studies) but not yet from independent groups.
• Cross-sectional — cannot establish temporal direction. Reduced state differentiation could be cause, consequence, or correlate of FM.
• Self-report measures of interoception have known reliability issues; the MAIA captures beliefs about interoception rather than interoceptive accuracy directly.
• The emBODY task is a body-drawing instrument, not a neural readout. The link to underlying computational/neural mechanisms (Strube's predicted precision-weighting deficits) is inferred, not directly measured.
• Sample size and demographics not in the abstract — need to check full text before citing prevalence/effect-size estimates.

Open questions raised

• Does emBODY classification accuracy correlate with TSPO-PET signal? With anti-SGC IgG titers? Which subtype does interoceptive deficit index?
• Do interventions targeting perceptual retraining, emotional state differentiation, or sensation re-interpretation reduce FM symptoms? (The paper proposes this; no trial yet exists.)
• Is the deficit specific to FM, or is it a general feature of nociplastic pain conditions (would test the framework's domain).
• What is the neural substrate of reduced body-map differentiation? (Connects to Hou 2026 thalamocortical decoupling and FPN-VAN alexithymia correlation.)

Triangulation notes

• Empirical anchor for interoceptive_inference: this is the missing FM-direct empirical evidence I flagged after Strube 2026 ingestion. The interoception entity now has a framework-tier paper (Strube) AND an FM-empirical paper (Herman).
• Closes part of Q13: predictive_coding_failure / interoceptive_inference is at minimum a real, measurable phenomenon in FM patients — not just a theoretical reframe. Whether it's distinct from CS or a generalization of it is still open, but the existence of the deficit is now established.
• Connects to Hou 2026: Hou found alexithymia uniquely correlated with FPN-VAN FC in FM. Herman shows alexithymia is elevated in FM. Together they begin to specify a circuit for the trait.
• Therapeutic implication: interoceptive-retraining interventions (perceptual retraining, emotional differentiation training, MBSR-style body scans done as differentiation practice rather than relaxation) become candidate non-pharmacological therapies with mechanism-anchored rationale rather than just empirical effect.

Bridges

• B4 (ME/CFS ↔ FM via post-viral neuroimmune) — strengthened. Strube 2026's framework was built around ME/CFS; if Herman's emBODY deficit replicates in ME/CFS, the mechanism-shared hypothesis gains evidence on two domains simultaneously.
• New bridge candidate: alexithymia ↔ interoception ↔ FM/ME/CFS network. The pattern of high alexithymia + poor interoceptive differentiation appears across these conditions and may share a circuit (FPN-VAN per Hou 2026). Worth flagging as B6 candidate.