NEW ENGLAND SEQUENCING AND TIMING (NEST)
Time: Saturday, March 13, 2004, 8:30 a.m. - 4:30 p.m.
Place: Lecture Room, Sterling Memorial Library,
Organizer: Bruno H. Repp, Haskins Laboratories
8:30 - 9:00
9:00 - 9:25
SHORT AND LONG TIME SCALE COMPONENTS OF RHYTHMIC MOVEMENT VARIABILITY
Richard C. Schmidt (College of the Holy Cross) and Theo Rhodes (University of Connecticut) E-mail: email@example.com
As technology for data acquisition has advanced, allowing behavioral data to be recorded for longer periods, researchers have been increasingly aware of the long-term serial dependence in their data records. Studies of timing tasks such as walking and tapping as well as other generic cognitive tasks have demonstrated self-similar, positive correlations in behavioral recordings that persist for the long term. A lesser-known fact is that such long-term memory processes (pink noise) are asymptotic in time (i.e., are only clearly revealed at the longest time scales), and variability at short-time scales may be differently structured (i.e., may contain not positive correlations but no correlations (white noise) or even negative correlations (blue noise)). Although past behavioral studies have found evidence for this bipartite temporal structuring, only minimal attention has been given to its implications. Using heuristic estimates of scaling exponents in tapping and wrist-pendulum swinging, the current research attempts to ascertain the difference in serial dependence of rhythmic timing for short and long time scales and determine how this serial structure depends upon the nature and properties of the timing task.
9:25 - 9:50
LIMBIC AND PARALIMBIC RESPONSES TO MUSIC PERFORMANCE: EFFECTS OF PERFORMANCE EXPRESSION AND MUSICAL TRAINING
Dinesh G. Nair, Edward W. Large, Kelly J. Jantzen, Fred L. Steinberg, and J. A. Scott Kelso (Center for Complex Systems & Brain Science, Florida Atlantic University) [presented by Ed Large] E-mail: firstname.lastname@example.org
We report two experiments using functional Magnetic Resonance Imaging (fMRI). Musicians and non-musicians listened to two versions of the same composition, one performed by a highly trained musician, and the other generated by a computer to conform as precisely as possible to the notated composition. Because the two stimuli were the same musical piece as defined by the musical notation, they shared all the basic features generally associated with music: pitch, melody, harmony, rhythm, grouping, meter, (mean) tempo, and architectonic structure. The two pieces differed only in those features associated with performance expression, which consist of local changes in timing and intensity - dynamics, articulation and rubato. The differences in brain activation while listening to the two versions of music index the neural processes associated with the perception of musical expression, but not with any features the two performances had in common. For both groups of subjects, increased activity in limbic association areas was observed for the expressive performance. We also observed significant differences between musicians and non-musicians in a number of brain regions.
9:50 - 10:15
METRICAL CATEGORIES IN INFANCY AND ADULTHOOD
Erin E. Hannon (Cornell University) and Sandra E. Trehub (U. of Toronto) E-mail: email@example.com
Long-term exposure to the musical structures of one's culture may result in lasting perceptual biases. Many findings indicate that North American adults are able to perceive and produce rhythmic patterns only when they have simple duration ratios and implied isochrony, which has led to the hypothesis that intrinsic biases constrain rhythmic and metrical structures in all human music. This claim is at odds with the prevalence of non-isochronous structures having complex duration ratios in music from many regions of the world, such as Eastern Europe, South Asia, Africa, and the Middle East. Thus, an alternative explanation is that the observed biases reflect greater familiarity with the simple ratios of Western music. Through prolonged exposure to Western metrical structures, adults may learn to rapidly categorize temporal intervals in a metrical context, an ability that would enable them to efficiently distinguish between structurally meaningful timing changes and expressive timing changes in individual performances. We presented North American and Bulgarian adults with simple and complex meter folk melodies from Eastern Europe. They rated the similarity of a set of altered versions of the original melodies that either preserved or violated the original metrical structure. Both groups of adults accurately rated structure-violating alterations as less similar than structure-preserving alterations for the familiar (Western) simple meter patterns, but only Bulgarian adults were able to distinguish the two types of alterations in complex meter melodies. When we presented the same stimuli to 6-month-old infants using a familiarization-preference procedure, they exhibited a novelty preference for structure-violating changes in both simple and complex metrical contexts. These findings suggest that adult biases may reflect long-term exposure to culture-specific metrical structure rather than intrinsic preferences for simple duration ratios and implied isochrony. Adults may have learned to disregard temporal information if it doesn't lead to an alternative interpretation typical of familiar metrical structures. To assess the developmental time course of such metrical category learning, an additional experiment is currently being conducted with 12-month-old infants. Preliminary data suggest that these older infants exhibit a novelty preference for structure-violating alterations only in simple meter contexts, indicating infants may learn culture-specific metrical categories some time in the first year. We interpret our findings as evidence of perceptual reorganization in music, with individuals becoming attuned to culture-specific metrical categories some time within the first year of life.
10:15 - 10:40
LATENT DIFFERENTIAL EQUATIONS MODELS OF INTERPERSONAL COORDINATION IN DANCE
Steve Boker, Pascal Deboeck, and Stacey Poponak (U. of Notre Dame) E-mail: firstname.lastname@example.org
When presented with a repeating rhythmic auditory stimulus, naive dancers who are asked to "move how they feel" tend to make relatively simple oscillating movements. We instructed one individual in dancing dyads to lead and the other individual to follow. In this paradigm, both dancers are presented with repeating auditory stimuli as well as repeating visual stimuli from observing their partner's movements. Dancers' movements were recorded using 8 sensors per individual and sampled at 80 Hz. We estimate the strength of coordinative coupling between the dancers using the Latent Differential Equations (LDE) approach and present the results of this analysis. LDE analysis is a recent method that allows modeling of the covariances between derivatives of one or more time series as a structural equations model. Of particular interest in these data are instances of nonstationarity when the dancers reorganize their movements from one pattern into another. We use a windowed extension to LDE in order to track time varying coefficients during these periods of nonstationarity.
10:40 - 11:10
11:10 - 11:35
THE CARTESIAN REFLEX PROJECT
John W. Moore, Robert J. Polewan (University of Massachusetts, Amherst), and Christopher M. Vigorito (Amherst College)
The Cartesian Reflex Project seeks to reestablish human eyelid conditioning as a paradigm for research and teaching in psychology and cognitive neuroscience. We shall illustrate the basic procedure for establishing anticipatory eyelid closure to a 'blink' command. The resulting CR, the Cartesian reflex, resembles eye blink conditioned responses observed with more conventional `unconditioned stimuli' such as an air puff (humans) or electric shock (rabbits). The 'blink' US serves the same purpose while side-stepping concerns about human subject well-being. This presentation discusses the wide range of applications of these methods in the domains of learning theory and affective/emotional expression. The heart of the project is a menu-driven protocol generator and low-cost hardware component that permits detection of eyelid movements. The system requires a computer, plug-in analog-digital converter, and recording potentiometer with power supply. Research aims include: (1) Investigations of basic parametric features of CSs, occasion setting, temporal uncertainty, etc. (2) Instrumental conditioning (e.g., incentives for criterion lid movements, devaluation, partial reinforcement). (3) Assessment of computational learning theories. (4) Tools for detecting reflexive deception in the tradition of Paul Ekman's research on facial expressions. Laboratories were once common features of Learning courses, but they have given way to Computer Rats in virtual worlds. We believe that students will be excited about performing and serving in real experiments, experiments which could well be significant platforms for research at the level of honors theses and up.
11:35 - 12:00
CONCURRENT TIMING OF AUDITORY AND VISUAL EVENTS IN HUMANS AND RATS Devin McAuley (Bowling Green State University) E-mail: email@example.com
Previous studies have shown that age-related impairments in divided attention and timing may involve the frontal cortex. In rats, lesions of the lateral agranular frontal cortex disrupt performance in a divided-attention variant of the peak-interval procedure referred to as a simultaneous temporal processing (STP) task (Olton et al., 1988). In the STP task, rats learn to associate auditory and visual cues with short and long fixed time intervals and then are probed on their ability to produce the duration of the cued target in focused and divided attention conditions. The present study examined the effects of age on a human version of the STP task and compared performance with rats. Eighty adults in three age groups (18Ð39, 60Ð74, 75+) and thirteen Fisher rats were tested on short and long target durations in focused and divided attention conditions. Target duration and attention condition were within subject manipulations. For half of the subjects, the short target was presented in the auditory modality and the long target was presented in the visual modality; for the other half of the subjects, the short target was presented in the visual modality and the long target was presented in the auditory modality. Overall, humans and rats showed a similar pattern of STP performance. For humans, there was no effect of age on averaged produced duration or on the variability of produced durations during focused attention testing. However, during divided attention testing, there was a large age-related increase in the variability of produced durations. In addition, the observed divided attention impairment was much more pronounced in the visual modality than in the auditory modality. Results will be discussed in the context of applying neurobiological studies of rat timing to the understanding of human timing.
12:00 - 12:25
TIMING, SEQUENCING, AND TEMPORAL PROCESSING RESOURCES
Scott W. Brown and Stephanie M. Merchant (U. of Southern Maine) E-mail: firstname.lastname@example.org
Many time perception experiments have employed dual-task methodology in which subjects perform a time judgment task concurrently with a nontemporal distractor task. Compared with single-task conditions, time judgments under dual-task conditions typically show more error and variability. Fewer studies, however, have examined performance on the distractor task. Distractor task performance is an important issue from an attentional resource viewpoint. Resource theory predicts that two concurrent tasks relying on common resources should produce a pattern of bidirectional interference. In the current research, we selected a sequence reasoning task to serve as a distractor task. We speculated that timing and sequencing both involve temporal information processing, and so probably invoke similar cognitive mechanisms. Therefore, simultaneous timing and sequencing tasks should produce mutual interference. Subjects performed timing and sequencing tasks both singly and concurrently in a series of 2-min trials. The timing task required subjects to generate a series of 5-sec temporal productions. The sequencing task required subjects to verify reasoning statements that described the ordering of a pair of letters (e.g., A precedes B-BA); subjects responded True or False to each statement as quickly as possible. Comparisons of single-task and dual-task conditions provided evidence of bidirectional interference: (a) the sequencing task interfered with timing by making temporal productions longer and more variable, and (b) the timing task interfered with sequencing by lengthening response times to the reasoning statements. The results suggest that timing and sequencing are related cognitive tasks that rely on a common set of temporal processing resources.
12:25 - 1:20
1:20 - 1:55
EVENT-RELATED POTENTIAL CORRELATES OF AUDITORY TEMPORAL ORGANIZATION
Joel S. Snyder (Rotman Research Institute, Toronto, Canada) E-mail: email@example.com
Observing brain dynamics in response to temporally structured patterns is helpful in understanding how we process and represent music, speech, and environmental sounds. Event-related potential (ERP) techniques provide a high temporal resolution tool for observing brain dynamics while people listen to patterns that evoke perceptually distinct states of auditory temporal organization. My colleagues and I have shown that early stages (10-75 ms post-stimulus onset) of auditory cortical processing respond with equal amplitude to sequences of sound events occurring at rates typical of music (0.3-8.0 Hz), whereas later stages (75-200 ms post-stimulus onset) show smaller responses at faster tempos. Early auditory responses show bottom-up and top-down representations of metrical accent during rhythmic stimulation. When tones in an alternating loud-soft pattern are occasionally omitted, brain responses occur as if the tone were present with larger responses for louder expected tones. Furthermore, paying attention to tones embedded in metrical patterns enhances the representation of metrical structure. More recently, our investigations have uncovered auditory processes related to segregating sound patterns into separate streams. Specifically, the time course and spectral content of neural activity undergoes reorganization when a sound pattern is heard as two streams rather than one. These studies suggest that rich representations of auditory temporal patterns arise at early stages of sound processing.
1:55 - 2:20
GAMMA-BAND RESPONSES TO PERTURBED AUDITORY SEQUENCES: EVIDENCE FOR SYNCHRONIZATION OF PERCEPTUAL PROCESSES
Theodore P. Zanto, Edward W. Large, Armin Fuchs, and J. A. Scott Kelso (Center for Complex Systems & Brain Science, Florida Atlantic U.) E-mail: firstname.lastname@example.org
We measured modulations of neuroelectric gamma-band activity (GBA) as subjects listened to isochronous pure-tone sequences with embedded temporal perturbations. Perturbations occurred every 6-10 tones, and at the locus of the perturbation, tones occurred early, on time, or late. In the absence of perturbations, induced (non-phase locked) GBA reached maximum power simultaneously with the occurrence of tone onsets, whereas evoked (phase-locked) GBA peaks were observed after onsets. During late perturbation trials, peaks in induced activity tended to precede tone onsets, and during early perturbation trials, induced peaks followed tone onsets. Induced peaks returned to synchrony after both types of perturbations. Early tones resulted in a marked increase in evoked GBA power at the locus of the perturbation. The latency of evoked GBA relative to tone onset, as well as some other features of the response, depended asymmetrically on the direction of the perturbation. The current results provide evidence supporting the role of GBA in temporal expectancy, and suggest that GBA is a good candidate for a neural correlate of pulse perception in musical rhythm.
2:20 - 2:45
CONTEXT EFFECTS IN SENSORIMOTOR TIMING AND COORDINATION
K. J. Jantzen1, F. L. Steinberg1,2, and J. A. S. Kelso1 (1Center for Complex Systems & Brain Science, Florida Atlantic University; 2University MRI and Diagnostic Imaging Inc., Boca Raton, FL) E-mail: email@example.com
Within the context of the continuation paradigm, neural activity patterns underlying the generation of internally timed movements reflect not only demands imposed by the ongoing task, but are also dependent on the method used to establish the required movement rate. Differences in neural activity that distinguish between pacing established using synchronized and syncopated coordination patterns persist during continuation such that substantially different networks are shown to support a single, simple motor behavior. We will review this recent fMRI work with particular emphasis on a recent study that investigated the degree to which this dependence is altered through the introduction of a delay between the pacing and continuation phase. Pacing was established by coordinating with an auditory metronome in either a synchronized or syncopated fashion. When the metronome ceased, subjects either continued moving without interruption, or stopped moving for 3, 6 or 9 seconds before being signaled to continue at the previously established rate, allowing for a parametric investigation of possible delay dependent effects. Two general patterns were observed. First, in SMA, dorsal premotor cortex, and cerebellum, the differences in BOLD intensity observed when comparing synchronized and syncopated pacing persisted during continuation. This was consistent for all delays used with no systematic dependence on the duration of delay. Second, in basal ganglia initial differences in neural activity did not persist during continuation. This was primarily due to an increase in BOLD amplitude during continuation following synchronization, with no change in continuation following syncopation. Again, there was no dependence on delay duration. These results demonstrate that the method of pacing and the pattern of neural activity it generates continue to influence the timing network even if movement execution is interrupted for a time span of several seconds, possibly reflecting a mnemonic mechanism for temporal interval representation.
2:45 - 3:15
3:15 - 3:40
SYNCHRONIZATION AND CONTINUATION TIMING IN TAPPING AND DRAWING MOVEMENTS
Howard Zelaznik and Breanna Studenka (Purdue University) E-Mail: Zelaznik@sla.purdue.edu
Over the past five years, we have argued that smooth drawing movements, such as circle drawing, with a temporal goal are timed differently than the timing seen in the traditional tapping task. We (Zelaznik, Spencer and Ivry, 2002, JEP:HP&P) have proposed that timing in tapping uses an event timing process, requiring a temporal representation, whereas timing in circle drawing is emergent. Turvey originally proposed that timing behavior is the result of the interaction between trajectory and neuronal dynamics. Timed behavior emerges from this interaction. Our work in support of this dichotomy has been derived from the continuation paradigm, in which we ascertain temporal consistency after the entraining timing signal has been disengaged. In the present study we directly compared timing during continuation with timing during synchronization for tapping and circle drawing at movement periods of 350, 400, 450, and 500 ms. We found that the synchronization trials exhibited the same linear relation between timing variance and interval duration squared as did continuation, for both tapping and drawing, although the intercept was higher in synchronization. When participants timed circles in synchronization with their non-dominant hand, the slope of the relation was negative. We interpret this striking finding as support for the notion that the emergent timing of a continuous movement requires specific practice.
3:40 - 4:05
MODELING CODRIFT IN SIMULTANEOUS ISOCHRONOUS FOUR LIMB TAPPING
Geoffrey L. Collier (South Carolina State University) and R. Todd Ogden (New York State Psychiatric Institute) E-mail: firstname.lastname@example.org
In prior research, we have presented a model that is an extension of that of Wing & Kristofferson, which decomposes the variance in isochronous tapping into two components: variance due to motor processes and variance due to a central clock. Our extension further decomposed the clock variance into a component due to timer drift and a residual, which we term the drift-free clock variance. This model did not look at the drift process itself. This drift process can be estimated by fitting a smooth function through the tapping time series. Any of a number of available smooth functions will approximate the drift process, but the statistically optimal function will have to take into account the specifics of the process, in particular, the lag 1 moving average (MA(1)) process induced by the motor delay. A model that does this is presented. As a side benefit to this model, estimators of the three sources of variance are obtained that have smaller standard errors than our earlier estimators, whose standard errors had been substantially larger than those of Wing & Kristofferson. Applications of the drift process estimators are suggested, such as studying attractor tempi.
4:05 - 4:30
MOTOR CONTROL IN SKILLED PIANO PERFORMANCE
Roger Chaffin, Bruce Kay, Theo Rhodes, and Michael Richardson (University of Connecticut) E-mail: email@example.com
"It is different each time I play" reports the pianist Emil Gilels. Can this be true? Performances by concert pianists in the Western classical tradition are normally highly prepared and polished; nuances of timing and dynamics are precisely replicated across performances. Is musical spontaneity really possible in such performances? Perhaps the differences are just a product of random noise in the motor system. On the other hand, if Gilels is right, it would seem that musically unpromising fluctuations can be damped down while those consistent with musical goals are allowed to persist (Latash, Scholz & Schöner, 2002). We compared 7 highly polished performances of J.S. Bach's Italian Concerto (Presto) by a concert pianist. Musically significant variations in tempo and dynamic level (musical gestures) were identified and compared across performances. Variation common to all performances accounted for 68% of the variation in tempo and 87% of the variation in dynamic level (loudness). Gilels' observation was confirmed by the finding that idiosyncratic differences in musical gestures between performances accounted for a small but significant portion of the variance (2% for tempo, 1% for dynamic level). Thus, performances varied more in the musical gestures that constituted the pianist's interpretation of the piece than at other points. According to the uncontrolled manifold (UCM) hypothesis, stability is achieved in skilled performance by allowing motor activity to vary more freely along less critical dimensions in order to achieve more precise control along more critical dimensions (Latash et al., 2002). In the case of musical performance it seems, somewhat counterintuitively, that performance is freest to vary at just those points where the artist has decided on a carefully calibrated variation in timing or dynamics in order to make a musical gesture. If this is right, then increased variability at musical gestures should be accompanied by decreased variability at points of control, places where the performance is more tightly controlled. We looked at places where the pianist reported paying attention during performance in order to ensure the accurate execution of critical movements such as the use of a particular fingering or a difficult jump. As predicted, there was less variation at the locations of these basic performance cues than at other locations in the piece. We conclude that repeated performances are different because movements are organized so that musical gestures can vary more freely than other aspects of the performance. The flexibility this provides allows the performance to retain the prized qualities of freshness and spontaneity at the same time that nuances of interpretation are executed automatically and reliably.
5:00 - 6:20
6:30 - 8:30