Age-pathology confounding is empirically demonstrated: suppressing age representation corrupts pathology representation in EEG foundation models.

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• age-pathology confounding, where it is impossible to suppress one concept without corrupting the otherquote0.881
  Verbatim description of age-pathology entanglement.
• Age-pathology confounding prevents independent steering of age and pathology concepts.claim0.875
  Interpretive assertion about clinical entanglement in the representations.
• Age-pathology confounding observed: impossible to suppress one concept without corrupting the other.finding0.873
  Empirical demonstration of entanglement between age and pathology features.
• age-pathology confoundingconcept0.855
  Entanglement phenomenon where age and pathology concepts cannot be independently steered without corrupting each other.
• EEG foundation models achieve state-of-the-art clinical performance yet their internal computations remain opaque, constituting a barrier to clinical trust.claim0.749
  Motivating claim for the entire paper
• We hypothesize that applying SAE-based mechanistic interpretability to EEG foundation models can expose representational failures and thereby improve clinical trust.hypothesis0.740
  Overarching motivating hypothesis of the paper
• Spectral decoder reveals pathological slow-wave suppression as a frequency signature of concept steering interventions in EEG foundation models.finding0.740
  Links latent space manipulation to known EEG neurophysiology
• Clinical concepts in EEG foundation models fall into three operational regimes: selectively steerable, encoded but entangled, and non-encoded.claim0.737
  Interpretive claim summarizing the spectrum of concept steerability discovered.