EEG covariance matrices live on the SPD manifold. Quantum states live on the Bures manifold. The question is whether these are the same space — and if so, what that means for how a quantum circuit reads neural signals.
Two patients' brain states trace opposite geodesics on the SPD manifold as seizure develops. The Fréchet mean (gold) is the population reference point. Polarity is the direction of departure from that point — and it is opposite for different patients.
A density matrix ρ represents a quantum state. For a pure state |ψ⟩, the density matrix is ρ = |ψ⟩⟨ψ|. Mixed states — partial knowledge of a quantum system — are weighted combinations of pure states.
Density matrices are always:
These properties are structurally identical to normalized covariance matrices. The trace constraint is the only addition.
When the A-Gate encodes an EEG window, the circuit's internal state at any moment is a density matrix — a point on the Bures manifold. The EEG covariance matrix that went in and the quantum state inside the circuit live in geometrically related spaces.
The Bures distance measures how distinguishable two quantum states are:
where F is the quantum fidelity:
Both manifolds are spaces of positive-semidefinite matrices with a natural Riemannian metric. The SPD manifold has no trace constraint; the Bures manifold normalizes to trace 1 (probability). That normalization is the only structural difference.
Geodesics — the shortest paths between points — have the same mathematical form on both manifolds. A covariance shift that moves one step along a geodesic on the SPD manifold corresponds to a proportional step on the Bures manifold after encoding. The geometry is preserved up to the normalization.
This is not an approximation. It is a consequence of the shared algebraic structure of both spaces.
Classical Riemannian methods project covariance matrices into tangent space, a geometry-aware linearization that preserves local structure. But linearization discards the global directional information: which way a seizure moves a patient’s brain state on the manifold, not just how far. That direction is polarity. Classical methods detect the magnitude of covariance shifts cross-patient, but the sign cancels when opposite-polarity patients are aggregated, explaining why Riemannian classifiers underperform despite using the right geometric framework.
The A-Gate resolves this because it is a fixed geometric probe. Its measurement basis doesn’t adapt to the data. When a standard-polarity patient’s covariance shifts one way relative to the circuit’s fixed axis, ⟨Z⟩ is positive; when an inverted-polarity patient shifts the opposite way, ⟨Z⟩ is negative. The sign is preserved through measurement precisely because the circuit doesn’t learn. A fixed instrument reading a directional quantity on a curved space.