Effects of music tempo upon submaximal cycling performance


Corresponding author: Professor J. Waterhouse, Research Institute for Sport and Exercise Sciences, Liverpool John Moores University, Henry Cotton Campus, 15-21 Webster Street, Liverpool L3 2ET, UK. Tel: +44 161 432 6352, Fax: +44 151 231 4353, E-mail: waterhouseathome@hotmail.com


In an in vivo laboratory controlled study, 12 healthy male students cycled at self-chosen work-rates while listening to a program of six popular music tracks of different tempi. The program lasted about 25 min and was performed on three occasions – unknown to the participants, its tempo was normal, increased by 10% or decreased by 10%. Work done, distance covered and cadence were measured at the end of each track, as were heart rate and subjective measures of exertion, thermal comfort and how much the music was liked. Speeding up the music program increased distance covered/unit time, power and pedal cadence by 2.1%, 3.5% and 0.7%, respectively; slowing the program produced falls of 3.8%, 9.8% and 5.9%. Average heart rate changes were +0.1% (faster program) and −2.2% (slower program). Perceived exertion and how much the music was liked increased (faster program) by 2.4% and 1.3%, respectively, and decreased (slower program) by 3.6% and 35.4%. That is, healthy individuals performing submaximal exercise not only worked harder with faster music but also chose to do so and enjoyed the music more when it was played at a faster tempo. Implications of these findings for improving training regimens are discussed.

As has been reviewed recently by Terry and Karageorghis (2006), music can have positive effects when played at the same time as physical activity is being performed. For example, music is often used by those taking part in recreational activity (Wininger & Pargman, 2003; Priest et al., 2004; Pazoki et al., 2007) and by patients recovering from a stroke or other cardiovascular problem and taking part in rehabilitation programs (Vollert et al., 2003; Metzger, 2004; Kim & Koh, 2005; Jeong & Kim, 2007; Mandel et al., 2007). Music has been used also by athletes during submaximal (Szmedra & Bacharach, 1998; Hayakawa et al., 2000; Potteiger et al., 2000; Nethery, 2002; Edworthy & Waring, 2006; Yamashita et al., 2006; Karageorghis et al., 2006a) and maximal (De Bourdeaudhuij et al., 2002; Atkinson et al., 2004; Tenenbaum et al., 2004; Crust, 2004a, b; Crust & Clough, 2006; Macone et al., 2006; Simpson & Karageorghis, 2006; Caria et al., 2007) exercise, as well as in their pre-competition preparations (Bishop et al., 2007; Eliakim et al., 2007), to improve performance.

The positive effects of music could arise in several ways. Music could enable a particular workload to be more acceptable and perceived as less arduous, but it could also mean that individuals chose to do more work without an increased sense of effort. These effects are often considered to indicate that music acts as a “distractor,” reducing the individual's perception of the work, fatigue and discomfort that are involved (Nethery et al., 1991; Szabo et al., 1999; Potteiger et al., 2000; De Bourdeaudhuij et al., 2002; Nethery, 2002; Crust, 2004b; Edworthy & Waring, 2006; Yamashita et al., 2006), or that it enables individuals to accept a higher level of effort and discomfort (Atkinson et al., 2004; Yamashita et al., 2006). In contrast, if the work-load is high enough then the individual's attention is directed to the painful effects of the exertion; attention cannot be focussed on the music and it becomes a negative distractor (Rejeski, 1985; Tenenbaum, 2001, 2005). For most people, however, the exercise intensity chosen is less than this (i.e., a positive effect of music would be expected), and it is such a lower intensity of exercise that has been investigated in the present study.

Several studies indicate that, particularly in an athletic context, fast tempi are preferable to slower ones (Szabo et al., 1999; Metzger, 2004; Priest et al., 2004; Tenenbaum et al., 2004; Crust & Clough, 2006; Edworthy & Waring, 2006; Karageorghis et al., 2006a; Simpson & Karageorghis, 2006). However, musical tastes differ and so the type of music that is most effective (see Karageorghis et al., 1999, 2006a, b; Wininger & Pargman, 2003) – or even if a rhythm alone, without lyrics or melody, acts as a substitute (Crust & Clough, 2006) – will depend upon the particular circumstance and the individual.

In the current study, we have attempted to reduce some of the variation that will arise due to interindividual preferences for particular pieces of music. We have studied participants on three occasions, in each case using the same set of six music tracks, which differed in duration and tempo. We have also, on the three different occasions, played the set of tracks (defined as the program) at different tempi (at the normal tempo, 10% above this tempo and 10% below this tempo), without changing the pitch of the music. Details of the duration of the tracks and the tempo of the music are given in Table 1.

Table 1.  Some details of the music tracks
Track NumberSlow programNormal programFast program
Duration (s)Tempo (beats/min)Duration (s)Tempo (beats/min)Duration (s)Tempo (beats/min)
  • *

    Tracks 1 and 6 were not used in the analysis (see “Methods”).

  • Includes pauses between tracks.

Total time (min:s)26:31 24:10 21:50 

This approach enabled us to investigate the effects of the duration and tempo of separate tracks, as well as of the overall program of music, upon physical activity and subjective responses to exercise of low-to-moderate intensity. That is, it was possible to determine the extent to which the tempo of the music or its length determined the workload voluntarily undertaken. The following predictions (hypotheses) can be made with regard to performance (see Table 1 for details of track tempi and length):

  • I.Predictions when comparing individual tracks (T2–T5):
  • A.If duration of track is important, then performance will vary between tracks:
  •   T2<T3 and T4;
  •   T2>T5;
  •   T3>T4 and T5;
  •   T4>T5.
  • B.If tempo of track is important, then performance will vary between tracks:
  •  T2<T3, T4 and T5;
  •  T3>T4;
  •  T3=T5;
  •  T4<T5.
  • II.Predictions when comparing the tempi of the programs (S, slow; N, normal; F, fast):
  • C.If overall duration of program is important, then performance will vary between programs:
  •  S>N and F;
  •   N>F.
  • D.If overall tempo of program is important, then performance will vary between programs:
  •  S<N and F;
  •  N<F.

It is these hypotheses that have been tested in the present study.



Twelve healthy male University students, without injury or cardiovascular disorders, participated; all lived conventional lifestyles, retiring 23:00–00:30 hours and rising 07:00–08:30 hours. None of the participants was either a smoker or taking any regular medication. Participants varied in their degree of physical fitness and amount of physical activity normally undertaken. Some anthropometric data, expressed as mean (SD), were age=21.3 (3.1) years, body mass=79.6 (13.3)  kg, height=1.786 (0.071) m.


Each participant was required to come to the laboratory (temperature 20.0±1.5 °C, relative humidity 70±10%) on four occasions – the “familiarization” day followed by the three study days. The 4 days were 1 week apart, and the visits took place at the same time of day (within 1 h) for any participant.

On the familiarization day, participants' anthropometric data were obtained and participants practized the protocol, including wearing and using the apparatus that would be used, understanding the measurements that would be made by the experimenter, and practising the subjective assessments that would be required. On this occasion also, participants chose a saddle height, the gearing of the cycle, and the music volume (played on a DVD player with headphones) that were comfortable to them (individuals differed in the volume they preferred but peak values were in the range 70–85 db). These values were recorded and used on the 3 days of the main study. The level of exercise was self-chosen, but each participant was told that it was one that he could maintain without strain for up to 30 min (the length of the longest program, see Table 1). The participants were told that the aim of the experiment was to investigate how reproducibly they could perform a bout of exercise on the three test occasions, and how reproducible were the effects produced by this exercise.

On this familiarization day also, any questions were answered. The participants then signed consent forms in the presence of a third, disinterested party. The protocol was approved by the University's Research Ethics Committee.

Participants were required not to undertake any non-habitual physical activity on the day before each of the three test days, nor to drink any alcohol during the evening before the test day. The protocol on the three test days was identical; it consisted of being seated for 30 min (while apparatus was attached and resting values were obtained) and then undertaking exercise for about 25 min. One aspect of the protocol that was not disclosed to the participants was that all music tracks on a particular day were played at the normal speed, at a speed 10% greater than this, or at a speed 10% lower than normal. The order of presenting the three speeds (fast, normal and slow) was randomized for the group as a whole.

Music tracks

Six music tracks were compiled specially for the study (Table 1). They reflected current popular taste among the undergraduate population, and had been chosen by one of the authors (J.H.), in consultation with colleagues, to form a balanced program. Whereas differences in basic tempo were fundamental to the study, and similarity of volume of the tracks was desirable (see Edworthy & Waring, 2006), the different lengths of the tracks was, from a scientific viewpoint, non-ideal. In practice, however, it was not possible to find a selection of music tracks that were of equal length as well as of comparable popularity.

Three discs were made: one at the normal (published) tempo, another with each track speeded up by 10% and the third with the tracks slowed down by 10%. The software used (Magix) did not change the pitch of the music. For all three discs, there was a 10 s interval between each of the tracks. After the three test days, participants were informed that different tempi that had been used on the three occasions. Some had thought that the music on some occasions seemed “brighter” than on others but, before being told, none had realized that the tempi had been altered on the three occasions of testing.

Apparatus and measurements

Heart rate was continuously measured by a monitor using electrodes attached to the chest (Polar, Kempele, Finland). The exercise took place on a cycle ergometer (Computrainer Pro Model 8002, Racer Mate, Seattle, WA, USA). It gave a continuous record of power output and distance covered. Participants chose the rate at which they pedalled (cadence) and also the force they exerted on the pedals, neither of which was controlled by the experimenters. No mention was made of whether or not participants should attempt to link their rate of pedalling to the tempo of the music.

Baseline heart rate was obtained during the last 10 min period before exercise. Responses to the physical activity were recorded after each of the six tracks. The total distance covered over the course of each track was recorded, and pedal cadence was calculated during the last 30 s of each track. In addition, mean heart rate and power output were recorded during the last 15 s of each track. During the 10-s gap between each track, participants were asked to state: (1) their perceived exertion, RPE, using a scale 6–20 (Borg, 1962); their thermal comfort, using a scale 1–9 (where “1” was “very cold,”“5” was “neutral” and “9” was “very hot”); and their opinion of the track just listened to (where “−5” was “disliked a lot,”“0” was “neither liked nor disliked” and “+5” was “liked a lot”).

Treatment of results

In order to remove effects of “getting going” and “nearly finished,” only results from tracks 2–5 have been analyzed.

Two-way analysis of variance (ANOVA) with repeated measures was used, the within-subject factors being track (four levels: tracks 2–5) and tempo of program (three levels: slow, normal and fast). Greenhouse–Geisser corrections were used, and significant differences within the main factors were assessed using Bonferroni's corrections. For correlations, the method of Bland and Altman (1995) was used, which takes into account multiple values being obtained from each participant.

The SPSS package, version 14, was used. Exact P-values have been given, results given as “0.000” in the statistics output being reported as “<0.0005.” Following convention, values where P<0.05 have been referred to as “significant” and those where 0.10>P>0.05 as “marginally significant.”


Effects of tempo

Figure 1 shows the total distance cycled during the four different tracks (tracks 2–5) when listening to the three programs at different tempi (slow, normal and fast tempi). When the average distances for the three tempi are compared, the relationships between the four tracks are those predicted by the hypothesis in Part I, section A. Moreover, the distances covered during the three different tempi support the hypothesis in Part II, section C for all four tracks. That is, the results indicate that the distance covered depended very much upon the duration of a music track.

Figure 1.

 The effect of track (T2–T5) and tempo of program (slow, S; normal, N; fast, F) on the total distance covered during each track.

Figure 2 shows the results for cadence, presented in the same way as in Fig. 1. A different result is evident. Thus, the relationships between the four tracks are closer to those predicted by the hypothesis in Part I, section B. In addition, the hypothesis in Part II, section D applies to all four tracks. That is, the results indicate that the cadence depended more upon the tempo of a music track.

Figure 2.

 The effect of track (T2–T5) and tempo of program (slow, S; normal, N; fast, F) on the cadence during each track.

All measured variables were investigated in the same way, enabling a decision to be made as to whether the duration or tempo of the track, and the duration or tempo of the program, exerted a greater influence. Table 2 summarizes the conclusions that could be drawn from these results, and Table 3 summarizes the statistical support for the differences observed. As indicated by Table 2, the total amounts of distance covered and work done, and the total numbers of pedal revolutions and heart beats, depended upon the duration of the track or program. In contrast, the rates measured during the program as a whole (distance/unit time, power, cadence and heart rate) depended upon the tempo of the music. Thus, when the tempo was raised by 10% above normal, then the mean increases in distance/unit time, power, cadence and heart rate were 2.1%, 3.5%, 0.7% and 0.1%, respectively; when the tempo was lowered by 10%, the mean decreases in distance/unit time, power, cadence and heart rate were 3.8%, 9.8%, 5.9% and 2.2%, respectively. Two results, cadence/music beat and heart rate/music beat, were the inverse of predictions for an effect of tempo (see hypotheses), because changes in cadence and heart rate were proportionally <10% change in music tempo.

Table 2.  Effects of track and tempo of program upon aspects of physical performance
VariableTrackTempo of programInteraction: Track × TempoInterpretation
F ratiodf* P-value
  • Whether effect due to duration of track or its tempo is indicated. Also shown are statistically significant interactions between track and tempo of program, and their interpretation.

  • *

    Degrees of freedom (Greenhouse–Geisser corrections).

  • Because change to cadence was less than the change in beat of the music.

  • Because change to heart rate was less than the change in beat of the music.

Total distanceDurationDuration7.343.0, 33.20.001Track 5, slow tempo, less than expected
Distance/Unit timeTempoTempo    
Total workDuration     
Total pedal revolutionsDurationDuration3.282.6, 28.20.042Track 5, slow tempo, less than expected
Cadence/Music beatInverse of tempoInverse of tempo4.632.2, 24.30.017Track 5, slow tempo, less than expected
Total heart beatsDurationDuration6.313.6, 39.10.001Track 5, slow tempo, less than expected
Heart rateTempo     
Increase in total heart beats above resting valueDuration     
Increase in heart rate above resting valueTempo     
Heart rate/music beatInverse of tempoInverse of tempo5.902.9, 31.40.003Tracks 3/5, slow tempo, less than expected
Thermal comfortTempo     
How much track liked Tempo    
Table 3.  Results from the ANOVA
VariableTrackDifferences†,‡Tempo of program
F ratiodf* P-value F ratiodf* P-valueDifferences†,§
  • Effects of track and tempo of program.

  • *

    Degrees of freedom (Greenhouse-Geisser corrections).

  • Values in bold, P<.05; in italics, 0.10>P>0.05; ns, P>0.10.

  • Differences between tracks (numbered 2–5) in effects produced when P<0.10 (Bonferroni corrections applied).

  • §

    Differences between tempo of programme in effects produced when P<0.10 (Bonferroni corrections applied); S, slow; N, normal; F, fast.

Total distance 350.9 1.3, 14.2 <0.0005 2<3/4; 2>5; 3>4/5; 4>5 22.4 1.2, 13.1 <0.0005 S>N/F; S>F
Distance/Unit time 11.47 2.3, 25.4 <0.0005 2<5; 3<5; 4<5 4.15 1.2, 13.3 0.056 S<F
Total work 47.91 1.4, 15.0 <0.0005 2<3; 2>5; 3>4/5; 4>50.78 1.6, 17.2 ns 
Power 6.72 2.2, 24.3 0.004 2<5; 4<5 3.83 1.6, 18.1 0.048 M<F
Total pedal revolutions 297.3 1.5, 16.7 <0.0005 2<3/4; 2>5; 3>4/5; 4>5 15.81 1.6, 17.1 <0.0005 S>F; M>F
Cadence 11.70 1.8, 20.0 0.001 2<3; 2>4; 2<5; 3>4; 4<5 5.21 1.4, 15.4 0.028 S<M; S<F
Cadence/Music beat 66.8 1.5, 16.3 <0.0005 2>3/4/5; 3<4; 4>5 17.02 1.4, 15.5 <0.0005 S>F; M>F
Total heart beats 399.2 1.2, 13.5 <0.0005 2<3/4; 2>5; 3>4/5; 4>5 23.48 1.8, 19.8 <0.0005 S>M/F; M>F
Heart rate 21.67 1.6, 17.9 <0.0005 2<3/5; 3>4; 3<5; 4<50.611.8, 19.8ns 
Increase in total heart beats above resting value 98.3 1.3, 13.9 <0.0005 2<3/4; 3>4/5; 4>5 2.422.0, 21.9ns 
Increase in heart rate above resting value 21.83 1.6, 17.9 <0.0005 2<3/5; 3>4; 3<5; 4<51.491.8, 19.4ns 
Heart rate/Music beat 180.2 1.4, 15.8 <0.0005 2>3/4/5; 3<4; 4>5 23.40 1.8, 20.1 <0.0005 S>M/F; M>F
RPE 14.62 1.5, 16.4 0.001 2<3/4/5; 3<5; 4<5 3.37 2.0, 21.9 0.053 S<F
Thermal comfort 30.47 1.4, 15.8 <0.0005 2<3/4/5; 4<5 0.901.6, 17.9ns 
How much track liked1.142.4, 26.3ns  3.80 1.8, 19.4 0.045 S<F

Subjective estimates (RPE, thermal comfort and how much tracks were liked) were influenced more by the tempo than the duration of the music (Tables 2 and 3). When the tempo was raised by 10%, the mean increases for the program as a whole in RPE, thermal comfort and how much the tracks were liked were 2.4%, 4.3% and 1.3%, respectively. When the tempo was lowered by 10%, the mean decreases in RPE and how much the tracks were liked were 3.6% and 35.4%, respectively. The comfort score rose 1.3% in this circumstance, and it is noteworthy that participants' scores on this scale (in the range “warm” to “hot”) indicated they were beginning to feel progressive thermal discomfort during the course of the program, regardless of its overall tempo.

The ANOVA indicated that there were comparatively few significant interactions between Track × Tempo of Program. Those present (Table 2) indicate that effects upon some aspects of physical performance and the responses to exercise when the music of track 5 was played at the slow tempo were less than would be predicted from a consideration of the other tracks and tempi.

Correlations between measured variables

Significant correlations between the measured variables are shown in Table 4. In the top part of this Table are shown the correlations using the four tracks played at the normal tempo. Significant positive correlations were found; that is, with a track that had a faster tempo, the rates at which individuals pedalled, performed work and covered distance, together with the physiological response to exercise (as inferred from the heart beat), all increased. There were few significant correlations between these variables and the extent to which the track being played was liked by the participants. The other subjective measures (RPE and thermal comfort) were generally positively related with the measures of activity and heart rate; that is, when they were performing more work and had a higher heart rate, participants felt that they were hotter and exerting more effort. The positive correlation between RPE and thermal comfort might reflect also the parallel upward trend shown by both these variables during the course of each program.

Table 4.  Correlations between aspects of physical activity
 PowerDistance/minHR↑ HRRPEThermal comfortLike
  • Top, results form all tracks, normal tempo. Bottom, separate tracks, three tempi. Method of Bland and Altman (1995) used.

  • *

    Significant correlations shown for each of tracks 2–5. All correlations positive except for that marked.

  • All correlations positive.+, P<0.05; (+), 0.10>P>0.05; O, P>0.10.

  • Emboldened, P<0.05; unemboldened, 0.10>P>0.05.

Cadence + + + + + (+)
Power  + + + +
Distance/min   + + + +
Heart rate (HR)     + +
↑ HR     + +
RPE      +
Thermal comfort      
Cadence 2,3,4,5 2,3,4,5 2,3,4,5 2,3,4,55  
Power  2,3,4,5 2,3,4,5 2,5 3,5 5  
Distance/min   2,3,4,5 2,3,4,5 3,4,5 5 
Heart rate (HR)     3,4,5 4,5 3,4
↑ HR     3,4,5  3,4
RPE      3,5 3,4
Thermal comfort       3,4,5*

Considering each track separately and investigating correlations when the program was performed at the three different tempi reduces effects resulting from the position of a music track in the program as well as the intrinsic tempo of the track. Results from this analysis (Table 4, bottom) indicate that significant or marginally significant positive correlations between the measures of activity and heart rate were generally found for all four tracks. Significant positive correlations between these variables and RPE were also generally present, except for track 2. The correlations with thermal comfort tended to be weaker and found most frequently with track 5. The extent to which the track being played was liked by the participants correlated positively with heart rate and the other subjective measures, but only for tracks 3 and 4.


The amount of physical activity performed overall (total distance covered, work done and number of pedal revolutions), as well as the physiological response to the exercise (total increase in the number of heart beats), increased with exercise duration, whether individual tracks or the tempo of the whole program was considered (Tables 2 and 3). However, to advocate extending the period of exercise, in order to improve the benefits of exercise, would often be impracticable; individuals generally train for a set period of time or a set distance.

More externally valid approaches would be to encourage as much exercise as possible in a set amount of time, or to advocate the performance of a set amount of exercise in as short a time as possible; the work rate would then be made as high as possible, subject to its acceptability to the participant, increasing the physiological and biochemical responses to the exercise. Attention became directed, therefore, toward one of the main aims of the study, the influence of the role of the tempo of the music upon the exercise variables (distance/unit time, power and cadence) and the physiological response (heart rate). These variables increased significantly when the tempo of the music was higher, whether individual tracks or the overall tempo of the program was considered (Tables 2 and 3). In other words, faster music increased the rate of working and the associated physiological response, whether the increased tempo was due to the intrinsic speed of the piece of music or having artificially speeded up the music track. This result, stressing the role of the tempo of the music, is in accord with those of Crust and Clough (2006), Edworthy and Waring (2006), Karageorghis et al. (2006a) and Priest et al. (2004), for example.

In the present study, increases in physical performance were accompanied by rises in the perceptions of effort involved (RPE) and heat load (Table 4). Moreover, participants tended to prefer the fast program (Table 3) and there were positive correlations between RPE, thermal comfort and how much the music was liked (Table 4, bottom). In other words, another of the main aims of the study – to investigate subjective responses to the tempo of the music – indicated that, when the music was played faster, the participants chose to accept, and even prefer, a greater degree of effort and what became, in effect, thermal discomfort. The observation that music was correlated with the performance of greater amounts of activity and acceptance of greater amounts of discomfort has been observed before (Atkinson et al., 2004). Other studies that have observed the same results have been interpreted to indicate that the music is acting as a “distracter” (Szabo et al., 1999; Potteiger et al., 2000; De Bourdeaudhuij et al., 2002; Nethery, 2002; Yamashita et al., 2006), and several authors have stressed the motivational role of music (Karageorghis et al., 1999; 2006b; Vollert et al., 2003; Wininger & Pargman, 2003; Crust, 2004b; Metzger, 2004; Tenenbaum et al., 2004; Macone et al., 2006; Simpson & Karageorghis, 2006). However, these studies have not investigated the role of tempo using the same pieces of music. It might be possible to assess these alternative interpretations (acceptance of more discomfort, distraction and motivation) if a comparison were made of the effects on performance and subjective responses when listening to music, comedy or instructional material.

The significant interactions between Track × Tempo (Table 2) and the correlation analyses (Table 4, bottom) indicate that the participants' responses varied between the different tracks. Performance during track 5 when the program tempo was slow was less than would have been predicted from a consideration of the other tracks; this finding might have resulted from a combination of lack of any “exhilaration” associated with the slightly slower tempo and increased thermal discomfort (see Crust, 2004b; Metzger, 2004; Crust & Clough, 2006; Edworthy & Waring, 2006; Karageorghis et al., 2006b). Volume, which has been found to play a role in other studies (Edworthy & Waring, 2006), is unlikely to have been important in the current study, because the tracks were recorded and played back at the same general volume.

As a further example of the differences between tracks, there were no significant correlations between the subjective variables for track 2 (Table 4, bottom). This might be because this track was too early in the program for factors such as thermal load and a sense of physical exertion and fatigue to be of importance. To investigate this would require the tracks to be played in a different order. It seems reasonable to infer from the results of Tables 3 and 4 that subjective factors are unimportant for short periods of exercise (track 2 had finished after about 6–7 min, see Table 1), and that the thermal load is more important with a longer bout of exercise (track 5 had finished after about 17–22 min). The relevance of this finding to exercise in hotter climates needs investigation. The effect of the tempo of individual tracks and the program as a whole was most marked for exercise bouts of intermediate duration (as indicated by tracks 3 and 4); at this point in the program, exercise had been continuing for some minutes and so effects of fatigue were beginning to be felt but thermal factors had not yet become dominant.


Faster music, whether due to the intrinsic tempo of the music or having increased the tempo artificially, enabled exercise to be performed at a greater work rate, and with a greater physiological effect and more positive subjective responses, than did slower music; these effects are due to some combination of motivational and distracting effects. These findings are relevant to those wishing to improve their training regimen – with faster music resulting in more work being performed in a set amount of time. However, our participants were healthy young males with no history of cardiovascular disorder, and they were exercising on a bicycle; whether our conclusions apply also to females, to other age groups and modes of exercise, and to those undergoing rehabilitation courses, cannot be inferred from the current results. Others (Priest et al., 2004; Karageorghis et al., 2006b, for example) have stressed interindividual differences when selecting an appropriate program of music to accompany exercise, and so further work is required to establish how generally applicable are our findings. Moreover, the current results apply only to exercise that is not of high-intensity; music need not improve performance when high work-loads are undertaken (Tenenbaum, 2005).