Method

Participants. Twenty students from the University of Lyon participated in Experiment 1. Musical practice, as measured by years of instrumental instruction, varied from 0 to 11 years, with a mean of 4.0 (SD = 4.1) and a median of 3.0.

Material. Twelve pairs of melodies were composed like the example shown in Figure 1. All melodies had a length of two bars of four beats each. The two melodies of a pair had exactly the same rhythmic patterns. The twelve melodic pairs had the same global rhythmic pattern: the first three beats of each bar consisted of eighth and/or sixteenth notes, the fourth beat was a quarter note, and the same rhythmic pattern was used for the two bars of each melody. The precise rhythmic figures used in a bar are shown in Figure 2. This diversity in rhythms for the three first beats was introduced to prevent the set of melodies from being too repetitive and boring for participants. The different rhythmic figures resulted in slight variations in the number of tones in a melody (M = 15.2, SD . The two melodies of a pair, in addition to having exactly the same rhythms, had almost the same tones. The two melodies of a pair differed only by one, possibly repeated, tone in the first bar. The serial position of this changed note was varied across melodic pairs from the second eighth note of the first beat to the fourth beat. This note change modified the tonal function of the target (the last tone of the melody, which was always a quarter note) to be either the tonic (related target) or the subdominant (less-related target). Twelve melodic pairs were composed so that each of the twelve major keys was represented twice: once with a melody ending on a tonic and once with a melody ending on a subdominant. Five additional melodies were used as examples during the task instructions.

Apparatus. Melodies were created in MIDI with Cubase SX2 software (Steinberg) and were transformed into audio files using The Grand (a VST piano instrument by Steinberg). MIDI velocity was constant for all pitches. Melodies were recorded at a tempo of 789.5 ms per beat without any expressive or stylistic timing variations. This tempo represents a duration of 789.5 ms for a quarter note, 394 ms for an eighth note, and 197 ms for a sixteenth note. The overall duration of a melody was 7000 ms (including the fade-out for the final target tone that increased its length to 1474 ms). Cubase’s micro-tuner was used to create the melodies with the last tone either in-tune or shifted in pitch by +/- 17 or +/- 35 cents (a cent is 1/100 of a semitone in logarithmic units). These two degrees of mistuning were selected from Warrier and Zatorre (2002) and represent a +/- 1% or 2% deviation in frequency. (A semitone represents a 5.946 % frequency difference.) The experiment was run on PsyScope software (Cohen, MacWhinney, Flatt, & Provost, 1993).

Procedure. The experiment consisted of a training phase and an experimental phase. In the training phase, the concept of “out-of-tune” notes was explained to participants with a familiar French song (“J’ai du bon tabac”) that was played first with its last tone in tune, and then with its last tone shifted in pitch by -35 cents, +35 cents, -17 cents, and +17 cents. The 35-cent pitch shifts were presented as “very out-of-tune” and the 17-cent pitch shifts as “slightly out-of-tune.” Participants were familiarized with the task with the five sample melodies. Participants had to indicate their response by pressing a key on the computer keyboard. If they judged the target tone to be in-tune, they pressed the “in-tune” key (“bonne”). If they judged it to be out-of-tune, they had to graduate their judgment on a 3-point scale: “slightly out-of-tune” (“légèrement fausse”), “out-of-tune” (“fausse”), or “very out-of-tune” (“très fausse”). In the experimental phase, participants judged the target tones of 120 melodies. The 12 pairs of melodies, which consisted of 12 related and 12 less-related melodies, were presented with the last note in-tune, 17 cents lower or higher, and 35 cents lower or higher; 12*2*5=120. The melodies were presented in a different random order for each participant. Participants had three seconds to answer, and then a 250 ms noise mask was presented. No feedback was given. Participants pressed another key to proceed to the next trial.

Data Analysis. Participants’ ratings were coded on a 0-3 scale, with 0 corresponding to “in-tune” judgments, 1 to “slightly out-of-tune”, 2 to “out of tune”, and 3 to “very out-of-tune”. Results were first analyzed with mean ratings averaged over the sequence set for each condition and participant (as in Warrier & Zatorre, 2002). We then analyzed discrimination performance between in-tune tones and each of the four mistuned conditions with areas under the receiver operating characteristic (ROC). Area under the ROC provides an unbiased estimate of discrimination performance where chance is 0.50 (Swets, 1973) and has the advantage of preserving more response information than d’ or A’, and of being uncorrelated with measures of response bias (unlike d’, see Dowling, Kwak, & Andrews, 1995; Verde, Macmillan & Rotello, 2006). Additionally, to assist in interpreting area scores, we analyzed hits and false alarms (with hit and false-alarm rates being percentages of “in-tune” responses to the five types of test stimuli).

Mean ratings and hits and false alarms were each analyzed with a 2 x 5 ANOVA, with Tonal Relatedness (Tonic vs. Subdominant) and Pitch Deviation (0, +17 cents, -17 cents, +35 cents, -35 cents) as within-participant factors. Area scores were analyzed with a 2 x 4 ANOVA, with Tonal Relatedness (Tonic / Subdominant) and Pitch Discrimination (0/+17, 0/-17, 0/+35, 0/-35 cents) as within-participant factors. The application of the Greenhouse-Geisser’s correction in these analyses did not change the significance patterns.

Figure 1. A)Example of the 12 pairs of melodies used in Experiment 1 and 2. B) The same melodies were used in Experiment 3, but with the last tone repeated. The tone differing between related and less-related melodies can be visually identified by the alteration marks. Examples of sound material are available at: http://olfac.univ-lyon1.fr/bt-sound.html.

Figure 2. Rhythmic figures used in the melodic set. For all melodies, the 2nd bar had exactly the same rhythm as the 1st bar, so only one bar is displayed. Of the 12 melodic pairs, 3 had the rhythmic figure A, 4 had the rhythmic figure B, 2 had the rhythmic figure C, 1 had the rhythmic figure D, 1 had the rhythmic figure E, and 1 had the rhythmic figure F.