SSVEP Spectrum Decoder Bench

Attended Target

Signal Parameters

Attention gain (SSVEP amplitude) 0.70

EEG noise level 0.40

Analysis window 2.0 s

Decoder Output

Decoded target: —
Confidence: 0%

Correct decodes

0 / 0

Decode rate

0.5 Hz

How it works

Four targets flicker at their own fixed frequency f_k. When attention locks onto target k, the simulated occipital (Oz) EEG channel picks up a steady-state visual evoked potential at f_k and its 2nd harmonic, plus faint crosstalk from the other targets and noise:

EEG(t) = G·[sin(2πf_k t) + 0.35·sin(2·2πf_k t)] + Σ 0.06·sin(2πf_j t) + N(t)·σ_noise

Unlike a single-bin check, this decoder sweeps the Goertzel algorithm across a continuous band (4–13 Hz) to draw the full power spectrum you see below the scope, then scores each of the 4 candidates by summing power at its fundamental and 2nd harmonic — a real harmonic-sum PSDA (power spectral density analysis) technique used to make SSVEP decoding more robust than reading the fundamental alone:

score(f) = P(f) + 0.5·P(2f)
decoded  = argmax_k score(f_k)
confidence = score(best) / Σ score(f_k)

A longer analysis window averages out more noise and raises accuracy, but a usable BCI speller needs to decode roughly every 1–2 seconds — the Analysis window slider makes that speed/accuracy trade-off explicit.

Flicker target (attended = filled) Simulated Oz EEG trace Power spectrum (decoded peak highlighted)