During the early stages of motor skill learning, performance of an unfamiliar task is uncoordinated and requires considerable cognitive effort in making decisions and executing movement (4,5,23,32). Novices must test untried solutions in their response repertoire in the attempt to discriminate between effective and ineffective response strategies (i.e., multiple degrees of freedom). However, with sustained practice, procedural memory evolves through a combination of implicit and explicit learning (2). Plasticity of the nervous system allows for reorganization of neural communication, and the development of an internal model via electrochemical adaptations and synaptic transmission among cortical and subcortical neurons. Such adaptations facilitate perceptuomotor integration, neuromuscular coordination, and the quality and consistency of motor skills. Consequently, lower-order skill components of the task, within the hierarchy of skilled components, demand less conscious regulation or executive control and performance of the task becomes more autonomous (1,5,8,24,31). For example, a novice target shooter must first learn the rules and strategies of competition, as well as master the fundamental shooting skills such as the postural stance, gun hold, aim, and timing of the trigger pull. Awareness of visual and kinesthetic feedback and the detection of errors require conscious thought and attention to technical points relevant to skill execution. However, with sustained practice, the individual becomes more familiar with the rules and strategies and is able to execute the fundamental skills of the task with greater automaticity, a state which functions to reduce task complexity and cognitive effort (5,8,24).
Few studies have investigated neurocognitive adaptations over extended periods of practice and instruction in real-world sport environments. In one of the few published longitudinal studies, Landers et al. (22) observed that novice archery students exhibited increased electroencephalographic (EEG) alpha power (12 Hz) over the left, but no changes over the right, temporal region in association with performance improvement after 14 wk of practice. Considering the well-established inverse relation between alpha power and cortical activation, this finding is consistent with an explanation of decreased cognitive effort (i.e., working memory) and development of an internal model as a result of practice, whereas the maintenance of right temporal activation would seem logical in light of the role of the right hemisphere in visuospatial processing, a primary component of the task. Research comparing expert and novice participants in target sports provides further support of this notion. Consistent findings have revealed that skilled performers exhibit greater alpha power (8–13 Hz) in the left temporal region during shot preparation (i.e., 5- to 8-s aiming period) relative to less skilled participants but manifest no differences in the right temporal region (11,14,28).
Synchronous oscillations of large assemblies of cortical neurons are an indication of increased interconnectivity and coherent organization among brain regions (19,25,27). As such, left temporal alpha synchronization may reflect inhibition of explicit working memory and a reduction in task-irrelevant brain processes. Attenuation of desynchronized cortical activity is important as desynchronized activity could potentially interfere with implicit procedural memory processes involved in sensorimotor integration (2,18). In light of the established role of the temporal regions in working memory and visual perception, and the relevance of such processes to the establishment of internal models to guide psychomotor activity, there is justification for the examination of the left and right temporal regions as done in previous marksmanship studies (9–12,14,17,22,28). However, much of the research that has examined cerebral cortical activity in relation to psychomotor skill has been cross-sectional (11,14,28) or has involved assessment of cortical change within single practice sessions. Additionally, the assessment of cortical activation in these studies has typically been limited and confined to the temporal and occipital regions. Therefore, further investigation is needed to examine cortical adaptations associated with practice over extended periods (i.e., weeks or months) and to examine a more broadly distributed spatial topography to determine the extent of practice-related change in cortical activity.
Accordingly, the present research examined regional (left and right temporal) as well as global cortical adaptations of first-year collegiate pistol shooters associated with practice over a time period of 14 wk and employed EEG signal processing techniques designed to enhance temporal resolution and sensitivity to task-related neurocognitive activity. A major advantage of EEG is that it provides a quantitative measure of electrical potential fluctuations of cortical pyramidal neurons with relatively high temporal resolution in contrast to other neuroimaging techniques. However, the resolution of EEG spectral estimates derived in previous research was limited to 500–2500 ms using fast-Fourier transform (FFT), perhaps concealing information about the dynamic changes in alpha power over the preparatory period. Application of alternative signal processing techniques such as event-related synchronization (ERS;26) and temporal spectral evolution (TSE;7) allow for higher temporal resolution of EEG spectral estimates. Therefore, a variation of ERS and TSE, event-related alpha power (ERAP), was implemented in the present study in order to increase the temporal resolution of alpha power estimates (125 ms), as was applied in a recent study of rifle marksmen (17). The broad alpha band (7–13 Hz) forms the peak of the EEG power spectrum, and its amplitude is attenuated in response to psychological demand in a regionally specific manner. However, greater specificity of cortical response to task-related demands can be achieved by isolation of higher alpha frequency components (alpha II; 11–13 Hz) from the lower frequency components (alpha I; 8–10 Hz) as the latter are more indicative of generalized arousal processes (11,17,19,27,31). Specifically, ERAP in the alpha II frequency band was extracted from response-locked EEG epochs because of its sensitivity to task-specific attentional processes.
Mean ERAP was predicted to increase with practice over time in the left temporal region as an index of increased functional organization (i.e., decreased reliance on working memory) during the sampled aiming period (−5 s to 0 s). Additionally, a higher rate of increase in ERAP (i.e., slope of power estimates over the 5-s aiming period) was predicted for the left temporal region with practice as a dynamic index of increased attentional focus during shot preparation. No such changes were predicted for the homologous right region because the visuospatial demands of aiming should remain relevant to task execution and, therefore, stable (9). Additionally, no changes were hypothesized for either of two control conditions (resting baseline, BL; and postural simulation, PS). Finally, exploratory analysis of ERAP mean power and slope was conducted over a spatially distributed topography to determine the extent of cerebral cortical changes associated with practice.
Thirteen (two female) right-handed participants who were ipsilateral eye dominant (age range 18–22 yr, mean = 19.08, SD = 1.50), with no shooting experience, were recruited from a pistol club at a United States Naval Academy. All of the participants completed a consent form that had been approved by the institutional review boards of both the military academy and the university with which the investigators were affiliated. Two individuals (one female) ceased participation with the club before completing their third test session. A sample size of 11 subjects was sufficient to detect a difference of 0.8 of the standard deviation of the differences in ERAP recorded at T3 from time 1 to time 3 (an effect size of 54% of the control) with 70% power when tested at the 0.05 level of significance. Therefore, statistical analyses were applied to data from the 11 remaining participants.
EEG preparation and acquisition.
A stretchable electrode cap (ElectroCap Inc., Eaton, OH) was fitted to the participant’s head in accord with standards of the international 10–20 system for recording EEG (15). The montage consisted of 11 sites (F3, Fz, F4, C3, Cz, C4, T3, T4, P3, Pz, and P4) referenced to the left mastoid online and re-referenced to digitally linked-mastoids offline. Electro-ocular activity (EOG) was measured with a bipolar electrode montage using two 10-mm diameter electrodes (Grass E5GH) attached superior and inferior to the orbital fossae of the right eye for vertical eye-movement (VEOG) and to the external canthi for horizontal eye-movement (HEOG). All EEG data were recorded and stored using Neuroscan software (Ver. 4.1) installed on a Gateway 2000 133-MHz Pentium computer with a Metrabyte 12-bit analog-to-digital converter. Amplifiers were calibrated for each participant separately through source input of a 50-μV, 10-Hz sinusoidal wave at all channels simultaneously. EEG signals were sampled at 256 Hz and amplified by a gain of 20,000 using a Grass Model 12A5 Neurodata Acquisition system with analog filter settings from 1 to 100 Hz. EOG activity was amplified 5000 times. Electrode impedance values were maintained from 5 to 10 kΩ at all sites. The trigger pull for each shooting trial was time-locked to the real-time continuous physiological recordings using an air pressure-sensitive event-marker. The pressure-sensitive sensor was attached just below the end of the gun barrel, which sent a voltage excursion to a designated channel of the EEG amplifier at the time of each shot.
Overview of daily practice sessions.
All practice was conducted at an indoor range at the military academy where the participants routinely practiced shooting a CO2-powered air pistol (Walther; Springfield, MA) during designated hours (3:30–5:00 p.m. or 5:00–6:30 p.m.). The range consisted of five shooting lanes, each with standard competition pistol targets (B-40) placed at a distance of 10 m. No telescopic or infrared pointers were used as aids for target aiming at any time. The participants practiced 5 d·wk−1 over the course of the study, and practice sessions were 60–75 min in length. Overall attendance for all participants was 87%. Group instruction by the coach was standardized for all participants at the beginning of each practice session (first 10–15 min). The latter 45–65 min were devoted to shooting practice, during which time the participants were required to complete 40–60 shots. The coach also provided performance evaluations, demonstrations, and verbal feedback to the participants routinely during all practice sessions throughout the course of the study. Additional individualized instruction was provided as required. The shooting positions of individuals varied slightly, but in general the standard position involved standing at the line of fire with the feet approximately shoulder width apart. From this position, the gun was raised in front of the shooter with the right arm (shoulder flexion along a vertical plane) while the left arm and hand were allowed to rest naturally along the left side of the body. Because competition rules prohibit the use of a second hand for stabilizing the gun, most of the participants tucked their left hand in their front left pocket or belt line or hooked their thumb through a belt loop. The trigger was operated by the right index finger. All of the participants wore safety glasses while shooting during all practice and test sessions. The left lens of all safety glasses was occluded by a strip of tape in order to minimize double vision and reduce facial muscle fatigue from squinting the left eye during aiming.
The participants practiced with air pistols exclusively from time 1 (beginning of practice period) to time 2 (midway through the practice period) and on alternate days between times 2 and 3 (end of practice period) (e.g., Mondays, Wednesdays, and Fridays). Between times 2 and 3, the participants integrated the use of 0.22-cal pistols (Walther) into their weekly practice routines (Tuesdays and Thursdays). The two instruments were essentially the same in terms of characteristics such as weight and trigger pressure. Shooting scores were not obtained on days when 0.22-cal pistols were used to practice the task. The number of days of air pistol practice and the associated shooting percentages were recorded over the 12- to 14-wk period. Performance for a given air pistol practice session was characterized as percentage accuracy by summing the point value of each target hit, dividing by the total points possible, and multiplying by 100. Concentric circles on the target ranged from 1 to 10, with 10 being the highest possible point total on any given shot (i.e., the bull). For example, if an individual scored 500 total points on 60 shots, his/her shooting percentage would be calculated as 500 pts/600 possible pts*100 = 83.3%. The coach provided reports of the scores obtained by the participants in all practice and test sessions to ensure expert and reliable scoring of the shooting percentages.
Overview of test sessions.
All test sessions were conducted at the same indoor range and designated practice hours using air pistol. Each of the participants was tested separately on three occasions beginning with an early-season measurement (time 1) and proceeding with mid- (time 2) and late-season (time 3) measures. Only two individuals could be tested on a given test day so that testing spanned a 1-wk period at each time of measurement. The duration of the study was 12–14 wk from time 1 to 3, depending on the sequential order of testing for a given participant, and intertest intervals were approximately 4–5 wk. Individual test sessions required approximately 75–90 min. For each of the three test sessions, the participants were first prepared for physiological recordings and then subjected to each of the conditions in the following order: 1) a 3-min baseline condition (BL), 2) a 10-trial postural simulation condition (PS), and 3) a 40-trial shooting condition (SH). EEG was recorded continuously during all three conditions. The order of assignment of the three conditions was fixed rather than random in an attempt to preserve ecological validity; this order accurately reflects actual practice and competition-related behavioral patterns of marksmen while a random order departs from conventional routines.
During BL, participants sat resting in a chair with eyes open, feet flat on the floor, and hands resting in their lap. They were instructed to sit quietly, look straight ahead, relax for a 3-min period, and to minimize movement and eye-blinks. During PS, participants assumed their standard shooting position with the pistol raised and pointed toward the target, but they were instructed not to align the sights on the target or pull the trigger. Rather, the experimenter manually triggered the event-marker on each trial (N = 10) after a stable hold period of 6 s by the participant. Thus, PS required postural demands similar to those for shooting but without precision aiming and preparatory motor demands for pulling the trigger. During SH, participants executed shooting trials (N = 40) from standard shooting positions after 5–10 sighting shots. A mean shooting percentage was calculated over the 40 trials for each of the test sessions at early (time 1), middle (time 2), and late (time 3) periods of the season.
ERAP was derived from 5-s data epochs using procedures similar to those for deriving event-related synchronization/desynchronization (ERS/ERD;19,25–27). Artifact-free epochs were band-pass filtered between 11 and 13 Hz, squared, and averaged over 125-ms intervals (32 samples). Single trials were ensemble-averaged, yielding a 5-s power-time curve with 125-ms temporal resolution (40 samples) for each site and each condition. Mean ERAP was then calculated by obtaining the grand average of all 40 points over the ensemble-averaged 5-s power-time curve for each site in each condition. In order to capitalize on the higher temporal resolution of the sampled power estimates, slope values were derived from a line-of-fit using linear regression to determine the rate of change of ERAP during the 5-s sampling periods (ERAP slope). For BL, 5-s sliding windows were used to derive ERAP with equal epoch durations. Horizontal (HEOG) and vertical (VEOG) electro-oculograms were recorded simultaneously with the EEG in order to identify and remove eye movement artifacts from EEG data. Eye movement artifacts were statistically removed from EEG channels using the eye artifact reduction algorithm (30) contained within Neuroscan software (Ver. 4.1). Any trials containing observable residual eye movement artifact, readily detectable by amplitude excursions in excess of 100 μV, were eliminated. The percentage of accepted trials for SH was 84.3%, 65.2%, and 73.6% for times 1, 2, and 3, respectively. For BL, the percentage of accepted epochs was 55.3%, 68.0%, and 55.6%, and for PS, the percentage of accepted epochs was 79.1%, 78.0%, and 80.0% for times 1, 2, and 3, respectively. All shooting scores were retained for all trials (none of the shooting scores were rejected). This strategy resulted in the highest possible stability of estimates of both shooting performance and artifact-free cortical activity as each was derived as an average of all available data (i.e., brain and behavior).
The primary analysis of ERAP mean and slope data used mixed model analysis of variance and covariance techniques (SAS, version 8.2). Mixed model techniques were used because the three factors of interest (three times, three conditions, and two sites) represent fixed sources of variability, whereas the among- and within-subject variability represents random sources of variability. Mixed model techniques also include features that allow the user to examine and account for variance heterogeneity and to appropriately account for correlated measures within a subject over time and/or among sites and conditions. The fixed portion of the model included sources of variation for time periods (1,2,3), conditions (BL, PS, and SH), and sites (T3 and T4) plus all possible two- and three-factor interactions of these factors. The number of practices per session was included as a continuous covariate to account for variation in the amount of practice among time periods. The random portion of the model included the variation among subjects and the interactions of subject with time, condition, site, and their two factor interactions. The residual variation was also included as a random effect. Mixed model techniques also include goodness of fit statistics that allow the user to assess how well the random portion of the model fit the residuals and to select the variance-covariance structure for the final model. Because the factors time, condition, and site are within-subject sources of variation, it was expected that these data were correlated and the goodness of fit indicated that this correlation is best accounted for by a form known as compound symmetry. Compound symmetry assumes homogeneity of variance and covariance for residuals among the levels of each factor. Besides compound symmetry, the unstructured matrix and the first-order auto regressive structures were also examined. Heterogeneous variance models were also considered for each factor. In addition to the above primary analysis (3 × 3 × 2) of ERAP mean and slope, an exploratory analysis of all 11 sites × 2 conditions (BL, SH) × 2 times (1,3) was also conducted. For this analysis, the contrast of time 1 versus time 3 was examined for each of the 11 sites during BL and SH conditions. Because modeling the data among 11 potentially correlated sites is very computer intensive, this analysis was restricted to the two most important conditions and the two most important times. Other features of this analysis are the same as those described above for the primary analysis.
The daily shooting percentages obtained during practice sessions were analyzed in two ways. First, daily percentages (daily%) were regressed on the number of practice sessions using linear regression in order to determine performance improvement with optimal temporal resolution. Second, the same daily percentages were averaged (mean season%) over early (up to time 1), middle (between times 1 and 2), and late (between times 2 and 3) periods of the season in order to examine the relation with mean ERAP, which was only examined during test periods corresponding to times 1, 2, and 3. For example, if on the first test session a participant had accumulated 10 d of practice, the mean of the 10 shooting percentages defined the individual’s level of performance achieved up to time 1 and the associated cumulative number of practice sessions was 10. If that same shooter then accumulated 25 d of practice by the second test session, mean performance at time 2 was derived from the scores obtained over practice sessions 11 to 25 and the associated cumulative number of practice sessions was 25, etc. Finally, shooting performance obtained during the three testing sessions, when EEG was recorded, was designated as test%. Mean season% and test% were analyzed separately by the same mixed model techniques described above. Each model included the effect of time period and number of practices per session to account for variation in the amount of practice among time periods. The designated random effects were due to subjects and the residual. The residuals were modeled to examine heterogeneity of variance among periods and to examine the correlations among repeated measures. Goodness of fit statistics were used to assess how well the random portion of the model fit the residuals to select the final model.
Simple regression of daily% on the number of practices revealed a significant linear increase over time (F (1,336) = 35.03, P < 0.01; see Fig. 1). The mixed model analysis of mean season%, accounting for variation in number of practices between time periods, also revealed a significant increase from time 1 to 3 (F (2,18.9) = 11.90, P < 0.01; see Table 1). These findings also validated that participants improved over the course of the season. However, when examination of performance was confined to that during the three test days only (test%), the mixed model analysis of test% revealed no change from time 1 to 3 (F (2,18.9) = 1.72, P = 0.21;Table 1).
ERAP of Temporal Sites
A significant condition by site interaction was observed for mean power (F (2,24.6) = 4.85, P < 0.02), as well as a significant main effect for condition (F (2,36.7) = 8.95, P < 0.01). No other interactions or main effects were significant, although the main effect for time approached significance (F (2,21.4) = 3.09, P < 0.07). Post hoc analyses of the condition by site interaction revealed that mean power was higher at T3 during SH compared with BL (F (1,36.0) = 2.67, P < 0.01) and PS (F (1,36.5) = 6.19, P < 0.02), whereas BL and PS did not differ from each other (P > 0.11). Similarly, mean power was higher at T4 during SH compared with BL (F (1,31.9) = 13.36, P < 0.01) and PS (F (1,32.3) = 13.36, P < 0.01), which also did not differ from each other (P > 0.36). The magnitude of the difference between T3 ERAP during SH relative to that observed at the same site during BL (ES = 1.32) and PS (ES = 0.80) was greater than the difference between T4 ERAP during SH and that observed at T4 during BL (ES = 0.85) and PS (ES = 0.63). Additionally, mean power during SH was higher at T3 than T4 (F (1,52.1) = 9.55, P < 0.01), whereas no differences were observed during either BL (P > 0.56) or PS (P > 0.16). Figure 2 illustrates the condition by site interaction.
Planned comparisons were examined to test the hypothesized simple effects that left temporal alpha power would increase from time 1 to 3 for SH, but not for either of the control conditions. The results indicated that mean power at T3 significantly increased in SH (P < 0.05) and PS (P < 0.03), but not in BL (P > 0. 23). Mean power at T4 did not change from time 1 to 3 in any of the three conditions (all P > 0.11). Figure 3 illustrates the changes over time for each site and condition
Analysis of slope values indexing the rate of change in power over 5-s sampling periods revealed a significant main effect for condition (F (2,22.8) = 3.91, P = 0.03). Figure 4 illustrates the magnitude of the slopes for each of the three conditions. No other effects were significant. Post hoc analyses of the condition main effect indicated that slopes were greater in SH (mean = 0.23, SE = 0.06) relative to both BL (mean = 0.08, SE = 0.06) and PS (mean = 0.03, SE = 0.06), which did not differ from each other. Planned comparisons of the hypothesized simple effects (i.e., slopes would increase from time 1 to 3 during SH, but not during either of the control conditions) were nonsignificant (P > 0.05). Figure 5 is a plot of ERAP over the 5-s aiming periods for sites T3 and T4 during each condition at times 1, 2, and 3 illustrating the ERAP points from which the mean power and slope values were derived.
Global Mean ERAP
Planned comparisons of the means between time 1 and time 3 during SH revealed significant increases at 8 of the 11 sites (F3, F4, C3, C4, Cz, T3, P3, and P4; t (21) ≥ 1.73, P < 0.05). The only change during baseline was limited to site C4 (t (21) = 2.13, P = 0.02). Figure 6 provides topographic maps based on all 11 sites of the ERAP estimates in 1-s intervals over the 5-s periods during SH and BL at times 1 and 3 to provide a more global representation of the cortical adaptations associated with practice.
Temporal Mean ERAP and Shooting Percentages
Linear and quadratic regression analyses of the relation between mean ERAP (T3 and T4) and shooting percentages (mean season% and test%) revealed significant quadratic relations. For each site and performance measure, increases in shooting percentages were associated with increases in mean ERAP up to an optimal level beyond which performance improvement leveled or deceased slightly (negative curvilinear). Figure 7 provides scatter plots with curvilinear lines of fit and second-order polynomial equations describing the relations.
Few training studies employing psychophysiological measures have been conducted to examine motor skill learning of a task beyond just a few hours of practice. To overcome this limitation, the present research examined the nature and time course of cortical adaptations associated with pistol shooting practice over a season. The shooting sports afford a number of benefits for the study of cortical dynamics during motor performance in light of the engaging nature of the task, ease of determination of performance improvement, and the motionless hold phase that allows for continuous EEG recording. Behavioral evidence of motor skill learning was supported in the present study by the observed increase in shooting percentages (daily% and mean season%) of the group over the 12- to 14-wk course of the season. The advantage of sampling performance on a daily basis was that it provided comparatively more accurate detection of change in shooting accuracy. The failure to detect such change during the three EEG test periods reflects the importance of sampling performance with higher frequency. As observed in a number of earlier reports (9–12,22,28), the present study also revealed asymmetry of temporal alpha power during shooting throughout all of the testing phases. This finding was indicative of relative right temporal activation during shooting along with higher power at both sites relative to the comparative conditions. Such a finding appears to be robust and implies that the task relies on specific neurobiological processes and relative synchronization in the left temporal region even in the early stages of skill acquisition. Further examination of mean ERAP at the temporal sites over the course of the training (i.e., time 1 contrasted to time 3) revealed that alpha power significantly increased in the left temporal region during shooting. In contrast, stability of cortical activation during shooting was observed in the right temporal region. The lack of change at T4 suggests engagement of the right temporal area likely due to the visuospatial processing associated with aiming (31).
The practice-related increase in mean ERAP that was observed for the left temporal region during the 5-s preparatory period for shooting was also observed in the postural simulation condition (although the highest mean levels of ERAP at times 1 and 3 were observed during shooting as shown in Figure 3) but not in the resting baseline condition. The emergence of such electrocortical change during both shooting and the postural control condition in a group who just learned the task suggests that experience with the basic task-related postural demands, by itself, contributes to fundamental changes in cortical activation. These postural-related changes, in coordination with the involved attentional processes, then appear to significantly influence the cortical dynamics during actual shooting or marksmanship. Other authors have, in fact, observed differences in cortical activation between postural control conditions and target shooting but the participants in those studies, unlike the present investigation, were highly skilled and had likely refined the cortical processes during aiming such that the EEG observed during shooting and control conditions were distinguishable. In addition, PS always preceded shooting in the present study, and it is possible that assumption of the shooting position may have primed the participants to adopt a psychological state that was substantively similar to that during actual shooting. In this manner, and in light of the level of skill development, the participants may have been mentally preparing to shoot resulting in similar practice-related neural activity.
The finding of increased left temporal ERAP with increased practice is consistent with the longitudinal study on archery students (22) and with the cross-sectional study between expert and novice rifle shooters (11). Landers et al. (22) attributed their findings of increased left temporal alpha power to changes in attentional strategies with practice. Similarly, Haufler et al. (11) interpreted their findings of higher alpha power in the left hemisphere sites in higher skilled shooters to indicate less explicit regulation of performance in the higher skilled shooters. Other researchers (17,28) have also reported an attention-related explanation for the patterns of left temporal alpha power observed during shooting. Both studies (17,28) revealed greater left temporal alpha power during shooting compared with motor control conditions that simulated postural and positional demands of shooting but without taking precise aim or releasing an actual shot. As stated above, the participants in these studies differed from those in the present in terms of expertise such that it is more likely that discrimination between cortical dynamics during positional control and shooting would emerge in experienced marksmen as compared with those who are relatively inexperienced. Considering that sensory selection or feature detection and response preparation are two neuropsychological components of attention (3), the increased left temporal alpha power in association with increased training experience observed in the present study and in previous research (22) suggests that relative synchronization in the left temporal region is related to goal-oriented attentional and memory processes. It is also noteworthy that a positive slope for ERAP was observed during the 5-s aiming period at all testing periods (i.e., early, middle, or late) of the marksmanship training, whereas no such increase in ERAP was noted at any testing phase during the other conditions. Such a finding during the aiming period of shooting reinforces earlier views advanced by Hatfield et al. (9), who also observed a progressive increase in T3 alpha power during aiming in elite marksmen, albeit with much less temporal resolution. Hatfield et al. interpreted their finding to mean that feature-detection and analytical process associated with the left temporal region are engaged early during the aiming period and become progressively reduced as the trigger pull is approached. The present results would suggest that this general pattern of cognitive engagement is maintained at various stages of skill or expertise but with less overall intensity as revealed by the concomitant mean increase in ERAP over the course of practice (i.e., 12–14 wk). The similar finding for ERAP slope at site T4 also implies a progressive reduction in the engagement of visuospatial processing over the course of the 5-s aiming period. Such an attentional strategy would make sense as the performer initially sights in on the target and then feels increasingly ready to execute the trigger pull.
Although intricately related to factors that influence attention, the role that memory subsumes in attentional processing and motor skill learning has not received much attention by previous researchers as an explanation for the higher levels of left temporal alpha power in higher skilled shooters. That is, greater left temporal alpha power observed during the moments preceding the trigger pull in more practiced and higher skilled shooters may reflect a decreased reliance on working memory processes for stimulus encoding of visual feedback. The lessened reliance on working memory may occur as a result of advanced declarative and procedural knowledge gained through the practice. The medial temporal lobes play an important role in the working memory system and are particularly involved in the encoding of new information and in the search for stored information in long-term memory (19). Practice and skill acquisition would likely reduce the reliance on working memory. In this manner, the proximity of site T3 to memory-related brain structures such as the hippocampal formation provides a plausible basis for the observed increase in alpha power with practice and experience. Further, object recognition and representation processes are organized by a semantic ventral pathway, which terminates in the temporal lobes while visuospatial orientation processes are organized by a pragmatic dorsal pathway, which terminates in the parietal cortex (16). The increased alpha power in the left temporal region observed in the present study may reflect increased functional coordination among these areas and the descending pathways responsible for the cortical control of movement. Results of the exploratory analysis, as illustrated by the topographic maps (Fig. 6), revealed that cortical adaptations were not just localized to the left temporal region but were spatially distributed. Such findings are consistent with a memory explanation because development of an internal model or procedural memory would involve the integration of many subprocesses and brain areas (e.g., visual, vestibular, proprioceptive, motor, and executive processes) and reflects increased stability in the system (18). Consistent with this notion of brain-related adaptation, a number of investigators have also reported that visual tracking is more stable in experienced shooters as indicated by longer quiet-eye periods (14), postural stability is greater (20), and the gun hold is more stable (21) than in less experienced shooters. In this manner, a number of “peripheral” processes associated with enhanced skill seem to reflect the change that occurs centrally.
From a broader perspective of motor skill learning, practice-related adaptations occur across multiple CNS subsystems, particularly the cerebral cortex, basal ganglia, and cerebellum (6). The neural network approach has also provided models based on other brain imaging methods (e.g., MEG, PET, and fMRI), indicating decreased activation in less relevant cortical areas and a shift toward greater reliance on cerebellar control with increased practice (2,6,13). At a neural level, a decrease in conscious working memory processes may facilitate stability in the motor processing centers of the brain by reducing potential sources of interference or competition (8,18). Existing evidence suggests that assemblies of neurons in cortical regions that are less salient to task execution exhibit decreased activation or increased synchronization as a result of practice (11,22,25,31). For specific regions of cerebral cortex that are most directly involved in the preparation and execution of movement, practice leads to a refinement in neural communication such that activation occurs only to the extent required by the task. That is, fewer but more highly organized networks may be used to represent a particular internal model via decreased contribution from cortico-striato-thalamic circuitry in providing feedback control and increased contribution from cortico-cerebellar loops as learning progresses (2). In the present study, greater synchrony was widespread across the cortex as training progressed, indicating a reduced role of cortical processes to accomplish the task. The implications are that motor processes are able to function with less variability and greater consistency if nonessential neural processes are minimized, which ultimately functions to improve the quality of movement (8). The increased alpha power in the left temporal region in the present study may be indicative of such minimization. Anecdotal evidence for a reduction in controlled processing or conscious regulation of the performance was provided by one of the participants in the present study. When asked to describe what, if any, experiential changes he had noticed from time 1 to time 3, the participant responded: “In the beginning, it was difficult to get equal daylight between the sights [before pulling the trigger]. Now I don’t even realize when I pull the trigger—it’s like wow the trigger went off?” By “equal daylight” the participant was referring to the process of centering the front sight (bar-shaped) with the rear sight (u-shaped) during the aiming period. His description of the differences in task demands from early to late season implied changes in sensorimotor processing during the preparatory period that were indicative of increased attentional automaticity in coordinating the timing of the trigger response with aiming mechanisms. In addition, the greater temporal resolution offered by the ERAP method indicated a linear increase of alpha power over the aiming period, whereas relative stability was observed in the comparative conditions. This finding would imply a more dynamic state of cortical activity that was manifest on all 3 testing days throughout the training experience. In terms of left temporal ERAP, the finding suggests greater reliance on explicit processes early in the hold and a remarkable shift away from such processing as the trigger pull approaches. In terms of right temporal ERAP, the finding suggests a greater reliance on and more effortful visuospatial processing that becomes less pronounced and less effortful as the trigger pull approaches. Apparently, this pattern of events is somewhat stable as skill acquisition proceeds.
Finally, curvilinear relations were observed between mean ERAP at both temporal sites (T3, T4) and both the long-term and short-term measures of performance (mean season%, test%) over the three test periods. These findings suggest that EEG indeed reflects task-related chronic neural adaptations and acute performance states, respectively. These findings are consistent with a fundamental principle of performance described by an inverted-U, indicating that better performance was associated with increasing synchrony of cortical activation to a point beyond which further synchrony was associated with a leveling or decline in performance. Previous researchers have reported a related finding such that heightened EEG alpha was in fact associated with poorer performance scores in archers (25,28) and a failure to execute or abort the shot in rifle marksmen (12).
A memory explanation of the EEG findings expands on previous interpretations that left temporal alpha power is related to attentional strategies by suggesting that memory functions may underlie the ability to selectively focus attention on task-relevant sensorimotor feedback information and response preparation without having to consciously think about it. Extended deliberate practice contributes to the store of task-relevant sensory and procedural information, which allows the experienced performer to more automatically access and use that information while negotiating the task (29,31). Such an explanation is also consistent with the schema theory for motor learning, which states that, as an internal model is modified and optimized, the role of feedback declines and there is a greater reliance on procedural memory or schema (31). From a cognitive information processing view, Hatfield and Hillman’s (8) psychomotor efficiency hypothesis predicts that sport-specific training and practice results in a higher work output (increased performance) with lower energy expenditure (decreased brain activation) resulting in greater consistency of the behavioral output or psychomotor performance.
The authors would like to acknowledge the contributions of Captain John Wilckens, Coach Cathy Callahan, Vice Admiral John Ryan, and members of the Midshipmen pistol club of the United States Naval Academy.
1. Badgaiyan, R. D. Executive control, willed action, and nonconscious processing. Hum. Brain Mapp. 9: 38–41, 2000.
2. Beek, P. J. Toward a theory of implicit learning in the perceptual-motor domain. Int. J. Sport Psychol. 31: 547–554, 1994.
3. Cohen, R. A. The Neuropsychology of Attention. New York: Plenum Press, 1993, pp. 3–545.
4. Cross, K. D. Role of practice in perceptual-motor learning. Am. J. Phys. Med. 46: 487–510, 1967.
5. Fitts, P. M., and M. I. Posner. Human Performance. Belmont, CA: Brooks/Cole, 1967, pp. 1–162.
6. Grafton, S. T., R. Hari, and S. Salenius. The human motor system. In: Brain Mapping: The Systems, A. W. Togaand J. C. Mazziotta (Eds.). San Diego: Academic Press, 2000, pp. 331–363.
7. Hari, R., R. Salmelin, J. P. Makela, S. Salenius, and M. Helle. Magnetoencephalographic cortical rhythms. Int. J. Psychophysiol. 26: 51–62, 1997.
8. Hatfield, B. D., and C. H. Hillman. The psychophysiology of sport: a mechanistic understanding of the psychology of superior performance. In: The Handbook of Research on Sport Psychology, 2nd Ed., R. N. Singer, H. A. Hausenblaus, and C. M. Janelle (Eds.). New York: John Wiley, 2001, pp. 362–386.
9. Hatfield, B. D., D. M. Landers, and W. J. Ray. Cognitive processes during self-paced motor performance: an electroencephalographic profile of skilled marksmen. J. Sport Psychol. 6: 42–59, 1984.
10. Hatfield, B. D., D. M. Landers, and W. J. Ray. Cardiovascular-CNS interactions during a self-paced, intentional attentive state: Elite marksmanship performance. Psychophysiology 24: 542–549, 1987.
11. Haufler, A. J., T. W. Spalding, D. L. Santa Maria, and B. D. Hatfield. Neuro-cognitive activity during a self-paced visuospatial task: comparative EEG profiles in marksmen and novice shooters. Biol. Psychol. 53: 131–160, 2000.
12. Hillman, C. H., R. J. Apparies, C. M. Janelle, and B. D. Hatfield. An electrocortical comparison of executed and rejected shots in skilled marksmen. Biol. Psychol. 52: 71–83, 2000.
13. Ingvar, M., and K. M. Petersson. Functional maps and brain networks. In: Brain Mapping: The Systems, A. W. Togaand J. C. Mazziotta (Eds.). San Diego: Academic Press, 2000, pp. 111–139.
14. Janelle, C. M., C. H. Hillman, R. J. Apparies, et al. Expertise differences in cortical activation and gaze behavior during rifle shooting. J. Sport Exerc. Psychol. 22: 167–182, 2000.
15. Jasper, H. H. The ten-twenty electrode system of the international federation. Electroencephalogr. Clin. Neurophysiol. 10: 371–375, 1958.
16. Jeannerod, M., M. A. Arbib, G. Rizzolatti, and H. Sakata. Grasping objects: The cortical mechanisms of visuomotor transformation. Trends Neurosci. 18: 314–320, 1994.
17. Kerick, S. E., K. Mcdowell, T. M. Hung, D. L. Santa Maria, T. W. Spalding, and B. D. Hatfield. The role of the left temporal region under the cognitive motor demands of shooting in skilled marksmen. Biol. Psychol. 58: 263–277, 2001.
18. Kinsbourne, M. Hemispheric specialization and the growth of human understanding. Am. Psychol. 37: 411–420, 1982.
19. Klimesch, W. EEG alpha and theta oscillations reflect cognitive and memory
performance: a review and analysis. Brain Res. Rev. 29: 169–195, 1999.
20. Konttinen, N., H. Lyytinen, and P. Era. Brain slow potentials and postural sway behavior during sharpshooting performance. J. Mot. Behav. 31: 11–20, 1999.
21. Konttinen, N., H. Lyytinen, and J. Viitasalo. Preparatory heart rate patterns in competitive rifle shooting. J. Sports Sci. 16: 235–242, 1998.
22. Landers, D. M., M. Han, W. Salazar, S. J. Petruzzello, K. A. Kubitz, and T. L. Gannon. Effects of learning on electroencephalographic and electrocardiographic patterns in novice archers. Int. J. Sport Psychol. 25: 313–330, 1994.
23. Lee, T. D., S. P. Swinnen, and D. J. Serrien. Cognitive effort and motor learning. Quest 46: 328–344, 1994.
24. O’Donnel, B. F., and R. A. Cohen. Attention: a component of information processing. In: The Neuropsychology of Attention, R. A. Cohen (Ed.). New York: Plenum Press, 1993, pp. 11–48.
25. Pfurtscheller, G. Event-related synchronization (ERS): an electrophysiological correlate of cortical areas at rest. Electroencephalogr. Clin. Neurophysiol. 83: 62–69, 1992.
26. Pfurtscheller, G., and A. Arannibar. Event-related cortical desycnchronization detected by power measurements of scalp. Electroencephalogr. Clin. Neurophysiol. 42: 817–826, 1977.
27. Pfurtscheller, G., and F. H. Lopes da Silva. Event-related synchronization and desynchronization: basic principles. Clin. Neurophysiol. 110: 1842–1857, 1999.
28. Salazar, W., D. M. Landers, S. J. Petruzzello, M. Han, D. J. Crews, and K. A. Kubitz. Hemispheric asymmetry, cardiac response, and performance in elite archers. Res. Q. Exerc. Sport 61: 351–359, 1990.
29. Schneider, W., and Shiffren, R. M. Controlled and automatic human information processing: I. Detection, search, and attention. Psychol. Rev. 84: 1–66, 1977.
30. Semlitsch, H. V., P. Anderer, P. Schuster, and O. Presslich. A solution for reliable and valid reduction of eye artifacts applied to the P300 ERP. Psychophysiology 23: 695–703, 1986.
31. Smith, M. E., L. K. Mcevoy, and A. Gevins. Neurophysiological indices of strategy development and skill acquisition. Cognit. Brain Res. 7: 389–404, 1999.
32. Starkes, J. L., W. Helsen, and R. Jack. Expert performance in sport and dance. In: The Handbook of Research on Sport Psychology, 2nd Ed., R. N. Singer, H. A. Hausenblaus, and C. M. Janelle (Eds.). New York: John Wiley, 2001, pp. 174–201.