A robust and self-paced BCI system based on a four class SSVEP paradigm: algorithms and protocols for a high-transfer-rate direct brain communication

Abstract

In this paper, we present, with particular focus on the adopted processing and identification chain and protocol-related solutions, a whole self-paced brain-computer interface system based on a 4-class steady-state visual evoked potentials (SSVEPs) paradigm. The proposed system incorporates an automated spatial filtering technique centred on the common spatial patterns (CSPs) method, an autoscaled and effective signal features extraction which is used for providing an unsupervised biofeedback, and a robust self-paced classifier based on the discriminant analysis theory. The adopted operating protocol is structured in a screening, training, and testing phase aimed at collecting user-specific information regarding best stimulation frequencies, optimal sources identification, and overall system processing chain calibration in only a few minutes. The system, validated on 11 healthy/pathologic subjects, has proven to be reliable in terms of achievable communication speed (up to 70 bit/min) and very robust to false positive identifications.

Figure 1

The stimulators placement and the graphic user interface layout.

Figure 2

Timings and structures of the adopted (a) screening and (b) training protocols.

Figure 3

An example of a joint time-frequency analysis (JTFA) calculated on a screening dataset varying the stimulation frequency in the 6–17 Hz range.

Figure 4

Conceptual scheme of the processing and identification chain for each of the four stimulation frequencies (s1, s2, s3, s4).

Figure 5

The adopted electrodes montage includes orange highlighted channels {T5,P3,Pz,P4,T6,O1,Oz,O2} using a linked mastoids reference and a ground channel placed in GND position.

Figure 6

Normalized weights associated to each electrode of the considered electrodes montage calculated using the Common Spatial Patterns method. Most of the subjects show a dominant activation localized over the O1-Oz-O2 electrode positions and, according to the weight signs, the optimal filter is usually a bipolar derivation possibly reinforced by parietotemporal activations.

Figure 7

Weights associated to each electrode of the considered electrodes montage using the manual channels combining approach described in Section 2.2.2. Positive weights refer to positive-contribution electrodes, while negative weights refer to negative-contribution electrodes.

Figure 8

Average results across all subjects (G1 + G2) obtained using the manual (blue line) and CSP-based (green line) methods with different lengths of the analysis window; (a) error rate, (b) delay time, and (c) bit-rate calculated using Wolpaw's definition.

Figure 9

Occurrences of each stimulation frequency of the screened 6–17 Hz range calculated on both G1 and G2 control groups.