Abstract
The clinical validity of four different low-redundant speech tests was calculated using four groups of 83 patients with retrocochlear or central auditory lesions. The speech tests used were: interrupted speech (7 or 10 interruptions/s) time-compressed speech and filtered speech. A comparison between patients and age-matched normal-hearing controls showed that the patients had significantly lower speech recognition score. The best sensitivity ratings of the tests were between 47% and 80%, the highest in patients with cerebellopontine angle tumours and temporal lobe lesions, and the lowest in vascular brainstem lesions. The speech tests with the highest sensitivity were 7 interruptions/s and time-compressed speech.