Efficient spatial filter approach permits single-trial decoding of speech items from M/EEG recordings