Noise-vocoding is a procedure that removes much of the spectral detail from speech, while preserving the temporal information (Shannon et al., 1995). Previous experiments have demonstrated that naïve English listeners show rapid perceptual adaptation to noise-vocoded sentences, and that this adaptation is driven by lexical information (Davis et al., 2005). However, these experiments did not manipulate information below the lexical level. The current study is the first to investigate the importance of speech rhythm as a cue to the segmentation of, and thus the perceptual adaptation to, noise-vocoded sentences. We hypothesized that, if language rhythm is a critical factor in adaptation to noise-vocoded speech, monolingual English listeners should receive a greater training benefit from noise-vocoded sentences in a foreign language that is rhythmically similar to English than from a language that is very different in rhythm. After a design by Pallier et al. (1998), we compared the training benefits of noise-vocoded sentences in English (stress-timed), Dutch (stress-timed) and Italian (syllable-timed) for subsequent comprehension performance on a set of English noise-vocoded sentences (we ensured that none of the participants in the Dutch and Italian conditions could understand the training language). Our prediction was that Dutch should give more successful adaptation than Italian, as Dutch belongs to the same rhythm class as English. The results are interpreted with additional reference to the role of individual variation in adaptation to distorted speech.