Experimental Brain Research

, Volume 225, Issue 2, pp 197–203

Skill-specific changes in somatosensory-evoked potentials and reaction times in baseball players

Authors

    • Institute for Human Movement and Medical Sciences
    • Department of Health and SportsNiigata University of Health and Welfare
  • Daisuke Sato
    • Institute for Human Movement and Medical Sciences
    • Department of Health and SportsNiigata University of Health and Welfare
  • Hideaki Onishi
    • Institute for Human Movement and Medical Sciences
    • Department of Physical TherapyNiigata University of Health and Welfare
  • Takuya Yoshida
    • Institute for Human Movement and Medical Sciences
    • Department of Health and SportsNiigata University of Health and Welfare
  • Yoko Horiuchi
    • Graduate School of Health and WelfareNiigata University of Health and Welfare
  • Sho Nakazawa
    • Graduate School of Health and WelfareNiigata University of Health and Welfare
  • Atsuo Maruyama
    • Institute for Human Movement and Medical Sciences
    • Department of Health and SportsNiigata University of Health and Welfare
Research Article

DOI: 10.1007/s00221-012-3361-8

Cite this article as:
Yamashiro, K., Sato, D., Onishi, H. et al. Exp Brain Res (2013) 225: 197. doi:10.1007/s00221-012-3361-8

Abstract

Athletic training is known to induce neuroplastic alterations in specific somatosensory circuits, which are reflected by changes in short-latency somatosensory-evoked potentials (SEPs). The aim of this study is to clarify whether specific training in athletes affects the long-latency SEPs related to information processing of stimulation. The long-latency SEPs P100 and N140 were recorded at midline cortical electrode positions (Fz, Cz, and Pz) in response to stimulation of the index finger of the dominant hand in fifteen baseball players (baseball group) and in fifteen athletes in sports such as swimming, track and field events, and soccer (sports group) that do not require fine somatosensory discrimination or motor control of the hand. The long-latency SEPs were measured under a passive condition (no response required) and a reaction time (RT) condition in which subjects were instructed to rapidly push a button in response to stimulus presentation. The peak P100 and peak N140 latencies and RT were significantly shorter in the baseball group than the sports group. Moreover, there were significant positive correlations between RT and both the peak P100 and the peak N140 latencies. Specific athletic training regimens that involve the hand may induce neuroplastic alterations in the cortical hand representation areas playing a vital role in rapid sensory processing and initiation of motor responses.

Keywords

PlasticitySEPsReaction time

Introduction

Motor skill training regimens are known to induce neuroplastic alterations in cortical areas associated with the sensory, motor, and cognitive tasks required by that regimen. For example, performance in sports is improved by training, which develops relevant sensory-motor skills, and these improved skills are reflected in neuroplastic alterations in relevant cortical regions. Studies on athletes using somatosensory-evoked potentials (SEPs) and event-related potentials (ERPs) following tactile stimulation suggest that specific training can modify the excitability of the somatosensory cortex and the neuronal circuits of the brain related to specific cognitive processes (Bulut et al. 2003; Iwadate et al. 2005; Murakami et al. 2008). Moreover, few magnetoencephalography studies have shown expanded somatotopic representations in the digit area of the primary somatosensory cortex (SI) in musicians (Elbert et al. 1995; Hashimoto et al. 2004).

Previous SEP and sensory evoked field studies (Bulut et al. 2003; Elbert et al. 1995; Hashimoto et al. 2004; Murakami et al. 2008; Thomas and Mitchell 1996) have mainly focused on shorter-latency components (e.g., N20 and P30 in upper limb and P37-N45 in lower limb) generated from SI. For example, Bulut et al. (2003) showed the latency of P60 following tibial nerve stimulation was significantly shorter in volleyball players than in sedentary males. In addition, Murakami et al. (2008) showed that the amplitude of P37-N45 following tibial nerve stimulation was significantly larger in football players than in non-athlete, and the amplitude of N20-25 following median nerve stimulation was significantly larger in racquetball players than in non-athletes. These studies suggest that long-term training induces neuroplastic alterations of neural circuit and excitation of SI. However, to our knowledge, no study has investigated the long-latency components P100 and N140 evoked by somatosensory stimulation in athletes (for review, see Nakata et al. 2010). Long-latency SEPs are associated with higher functions such as selective attention and spatial attention (Eimer and Forster 2003; Kida et al. 2004, 2006a; Waberski et al. 2002). In fact, Tanaka et al. (2008) reported that late activity (>80 ms) may reflect perception and integration of sensory information from earlier stages. In addition, only few electroencephalography (EEG) studies on somatosensory systems have shown that P100 and N140 are involved in passive attention (Kekoni et al. 1996; Kida et al. 2004, 2006a) and automatic change detection (Yamashiro et al. 2008). On the basis of these findings, long-latency SEPs should be related to behavior such as reaction time (RT) and/or sensory discrimination including decision making.

We hypothesized that if specific training alters sensory processing between the trained limb(s) and associated cortex, sensory perception and decisions made on the basis of these perceptions should be faster in a trained group than in an untrained group. To test this hypothesis, we compared long-latency SEPs and RTs in response to index finger stimulation of the dominant hand in baseball players (baseball group) with those in athletes involved in other sports (such as track and field events, soccer, and swimming) that do not require the same refined hand sensation for training or performance (sports group). We expected that long-latency SEPs and RTs would be faster in the baseball group than the sports group.

Methods

Subjects

SEPs were measured in 30 healthy volunteers (28 males and 2 females). Subjects were university students and graduate university students. Fifteen subjects had played baseball for more than 9 (mean 11.4 ± 1.7) years (baseball group), whereas 15 subjects had played other sports such as track and field events, swimming, and soccer for more than 6 (mean 9.3 ± 2.9) years (sports group). The baseball group and sports group were matched for age (mean, 20.3 ± 1.1 vs. 21.7 ± 2.9 years), height (172.2 ± 6.3 vs. 171.1 ± 6.9 cm). The present study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Niigata University of Health and Welfare, Niigata, Japan. Written informed consent was obtained from all subjects.

Somatosensory stimulation

SEPs were elicited by a constant current square wave pulse (pulse duration 0.5 ms) delivered to the index finger of the dominant hand. The intensity of the stimulus was three times the sensory threshold for baseball group 1.55 ± 0.52 mA and for sports group 1.44 ± 0.29 mA and was not reported as painful. The inter-stimulus intervals (ISIs) were randomized between 5 and 8 s.

Conditions

SEPs were recorded under passive and reaction time (RT) conditions. In the passive condition, subjects were instructed to ignore the stimulus and not blink for 1 s after the stimulation. In the RT condition, subjects were instructed to press a button using the index finger of the dominant hand as fast as possible when they perceived the somatosensory stimulation.

Recording

SynAmps amplifier system and scan 4.3 software (Neuroscan, El Paso, TX, USA) were used during EEG. The EEG was recorded using 9 scalp electrodes placed at F3/F4, C3/C4, P3/P4, Fz, Cz, and Pz according to the 10–20 system. The left earlobe was used as a reference. Electrode impedance was maintained below 5 KΩ. EEG signals were recorded with a notch filter (50 Hz) at a sampling rate of 1,000 Hz. Trials exceeding ±100 μV were excluded from averaging. For both passive and RT conditions, 70 artifact-free trials were averaged for each subject. The temporal window of analysis was from 50 ms before to 300 ms after the stimulus onset. The 50-ms period before the stimulus was used as the baseline. Band-pass filtering was set at 0.5–70 Hz.

Statistical analyses

Electrical stimulation elicited the P100 and N140 SEP components in all conditions and groups. Since P100 and N140 are maximal over the midline electrodes, the present study analyzed only signals among the midline electrodes (Fz, Cz, and Pz) and the reference. The peak amplitudes of P100 and N140 were measured relative to the pre-stimulus baseline. The peak latency and amplitude of P100 and N140 at Fz, Cz, and Pz were measured between 50–120 and 100–190 ms, respectively. They were subjected to a three-way analysis of variance (ANOVA) with two within-subject factors, condition (passive vs. reaction time) and electrode positions (Fz vs. Cz vs. Pz) and one between-subject factor, group (baseball vs. sports). The Greenhouse-Geisser epsilon was used to correct the degrees of freedom. Post hoc tests (Bonferroni) were performed to assess the pairwise differences in the peak SEP amplitude and latency between the groups, conditions, and electrode positions. In addition, differences in RT were tested by independent t tests. Statistical significance was set at p < 0.05. Moreover, we analyzed the bivariate correlations between the RT and P100 and N140 latencies.

Results

P100

Figure 1 shows the grand-averaged waveforms at the nine electrodes for the baseball group and the sports group under passive and RT conditions. The peak latency and amplitude of P100 at three midline electrode sites (Fz, CZ, Pz) for both groups and conditions are shown in Table 1. Three-way ANOVA revealed a significant difference in P100 latency between groups (F(1,28) = 19.070, p < 0.001) and between conditions (passive vs. RT) (F(1,28) = 4.574, p < 0.05). Latencies were not significantly different between the electrodes, and there was no significant condition–electrode interaction. The P100 peak latency under both the passive and the RT conditions was significantly shorter in the baseball group compared with the sports group (Table 1). In addition, the P100 peak latency under RT condition was significantly shorter than that under passive condition. The three-way ANOVA revealed a significant difference in peak P100 amplitude between the two conditions (F(1,28) = 28.349, p < 0.001) and a condition–electrode interaction (F(2,56) = 9.9, p < 0.001, ε = 0.748) but not between groups (baseball vs. sports). The peak P100 amplitude was significantly larger under the RT condition compared with the passive condition in both groups. Post hoc tests also showed that the peak P100 amplitude at Cz was significantly larger than that at Pz in the passive condition.
https://static-content.springer.com/image/art%3A10.1007%2Fs00221-012-3361-8/MediaObjects/221_2012_3361_Fig1_HTML.gif
Fig. 1

Grand-averaged somatosensory-evoked potential waveforms at nine cortical electrodes for the baseball group and the sports group under passive and reaction time conditions

Table 1

The peak amplitude and latency ± SD of P100 and N140

 

P100

N140

Sports group

Baseball group

Sports group

Baseball group

Passive

Reaction

Passive

Reaction

Passive

Reaction

Passive

Reaction

Latency (ms)

Fz

98 ± 13

97 ± 13

85 ± 13

79 ± 13

154 ± 10

148 ± 10

133 ± 20

134 ± 18

Cz

99 ± 13

95 ± 10

81 ± 12

77 ± 11

154 ± 13

147 ± 16

134 ± 21

131 ± 16

Pz

100 ± 13

96 ± 9

81 ± 16

81 ± 12

156 ± 17

155 ± 15

134 ± 22

134 ± 21

Amplitude (μV)

Fz

4.1 ± 2.7

5.8 ± 1.9

5.1 ± 4.1

5.5 ± 4.1

−6.5 ± 4.1

−8.6 ± 3.3

−5.7 ± 6.6

−9.4 ± 5.2

Cz

4.1 ± 2.2

6.5 ± 3.0

5.4 ± 4.6

6.7 ± 4.2

−-6.5 ± 3.6

−9.1 ± 4.2

−4.5 ± 4.4

−8.3 ± 3.3

Pz

3.5 ± 2.0

6.7 ± 2.5

4.4 ± 3.8

6.2 ± 3.6

−4.6 ± 2.9

−6.1 ± 3.4

−3.5 ± 4.1

−5.1 ± 4.0

N140

The three-way ANOVA revealed a significant difference in the latency of N140 between the groups (F(1,28) = 12.227, p < 0.01) and between conditions (passive vs. RT) (F(1,28) = 4.624, p < 0.05). Latencies were not significantly different between the electrodes, and there was no significant condition–electrode interaction. The peak latency of N140 in the baseball group was significantly shorter than that in the sports group (Table 1). The three-way ANOVA revealed a significant difference in peak N140 amplitude between the two conditions (F(1,28) = 16.734, p < 0.001) and a condition–electrode interaction (F(2,56) = 8.142, p < 0.001, ε = 0.895) but not between groups (baseball vs. sports). The peak N140 amplitude was significantly larger under the RT condition compared with the passive condition in both groups. Post hoc tests also showed that the peak N140 amplitude at Fz and Cz was significantly larger than that at Pz in the passive and RT conditions.

The relationship between long-latency SEPs and RT

The independent t test revealed a significant difference in RT between the groups (p < 0.001). RT in the baseball group was significantly shorter than that in the sports group (Table 2). When the relationship between the P100 latency and RT was analyzed across all subjects, there was a significant positive correlation at the three midline electrodes (Fig. 2a; Table 3). Additionally, there was a significant positive correlation between the N140 latency and RT at Fz and Cz (Fig. 2b; Table 3).
Table 2

The handedness, reaction times, and the latency of P100 and N140 at Cz under RT condition in each subject

Subject no

Baseball

Sports

Handedness

P100

N140

RT

Subject no

Handedness

P100

N140

RT

1

Right

68

121

218

1

Right

103

151

224

2

Right

68

143

192

2

Right

86

144

205

3

Right

79

134

192

3

Right

94

125

199

4

Right

78

130

205

4

Right

83

142

201

5

Left

77

116

214

5

Right

111

143

237

6

Right

91

141

199

6

Right

98

185

249

7

Right

59

128

189

7

Right

103

150

226

8

Left

54

104

182

8

Right

97

141

218

9

Right

83

130

170

9

Right

99

155

265

10

Right

90

140

218

10

Right

92

141

214

11

Right

80

123

171

11

Right

93

161

209

12

Right

89

148

219

12

Right

104

147

221

13

Right

79

154

187

13

Right

100

159

240

14

Right

72

101

199

14

Right

72

112

165

15

Right

83

145

208

15

Right

86

146

189

Mean

 

77

131

198

Mean

 

95

147

217

SD

11

16

16

SD

10

16

25

https://static-content.springer.com/image/art%3A10.1007%2Fs00221-012-3361-8/MediaObjects/221_2012_3361_Fig2_HTML.gif
Fig. 2

Correlation between the reaction time (RT) and the latency of peak P100 at electrode Cz and between RT and latency of peak N140 at Cz

Table 3

The r values of correlations between reaction time and the latencies of P100 and N140 at the three midline electrodes

Latency

Fz

Cz

Pz

P100

0.562**

0.666**

0.565**

N140

0.449*

0.603**

0.290

p < 0.05; ** p < 0.01

Discussion

Skill training can alter SEPs by inducing neuroplastic alterations in the specific cortical areas involved in the training. In this study, we tested whether specific training modifies somatosensory processes related to the long-latency potentials P100 and N140. We compared baseball players with athletes involved in other sports (soccer, swimming, track and field events) that do not require fine sensory perception and motor control of the hands and fingers. The results showed that (a) the peak latency of P100 and N140 latencies and RT in response to index finger stimulation were significantly shorter in the baseball players than that in other athletes and (b) there were significant positive correlations between RT and both the peak latencies of P100 and N140. These findings suggest that specific training alters higher-level perceptual processing in the somatosensory system.

Neuroplastic alterations in somatosensory cortex have been demonstrated in athletes and other professionals requiring extensive specific sensorimotor skill training (Bulut et al. 2003; Elbert et al. 1995; Hashimoto et al. 2004; Thomas and Mitchell 1996). Bulut et al. (2003) showed that the latency of P2 following tibial nerve stimulation was significantly shorter in volleyball players than in sedentary males. In addition, Iwadate et al. (2005) showed that long-term training shortens the P300 latency in soccer players compared with non-athletes, suggesting that long-term training alters both somatosensory processing in SI and cognitive processing. In contrast, Thomas and Mitchell (1996) showed that the latencies of P9, P11, P13/4, N20, and P25 following median nerve stimulation were not significantly different between runners, gymnasts, and non-athletes, suggesting that athletic training may have little impact on early sensory processes reflected by short-latency waveforms.

In the present study, the latencies of P100 and N140 and RT were shortened by approximately 20 ms in the baseball group compared with the sports group, which was consistent with previous findings that long-term specific training can alter somatosensory pathways and cognitive processes (Bulut et al. 2003; Iwadate et al. 2005). We suggest that the differences in results among previous studies are attributable mainly to the specific neural structures activated during training.

Baseball players require superior somatosensory processing in the cortical hand area for catching, batting, and throwing. Similarly, a swimmer requires fine motor control under the water. However, baseball players have a greater opportunity to grasp the ball and bat; thus, we speculate that grasping movement facilitates neuroplastic alteration. Specific cortical grasping network has been reported by a previous study in monkeys and humans (Cavina-Pratesi et al. 2010; Nelissen and Vanduffel 2011). The candidates for this network include primary and secondary somatosensory cortex, and this area is consistent with somatosensory processing circuit (Inui et al. 2004). Therefore, long-term training for baseball might specifically enhance this circuit. In contrast, soccer players and runners do not require fine discrimination of hand sensations, and it has been demonstrated that training-induced neuroplastic alterations occurred only in the trained limb (Elbert et al. 1995; Hashimoto et al. 2004; Iwadate et al. 2005; Murakami et al. 2008; Schwenkreis et al. 2007). Moreover, it has been reported that intensive physical exercise affects the amplitude of auditory and visual P300 (Polich and Lardon 1997). A recent study of auditory evoked potentials (AEPs) has shown enhancement of N1c and P2 auditory responses in skilled musicians compared with non-musicians (Shahin et al. 2003) and an effect of training on P2 enhancement in adult and child musicians compared with non-musicians (Trainor et al. 2003). Therefore, it is possible that the quality and amount of training affect neuroplastic alterations in the brain.

In the visual modality, Delpont et al. (1991) have reported that the latency of visual P100 in tennis players was significantly shorter than that of visual P100 in control subjects and rowers. They suggested that tennis players must greatly develop their ability to rapidly process sensory information; therefore, shortening of visual P100 in tennis player could be related to this ability. Similarly, Ozmerdivenli et al. (2005) have reported that the latency of visual N145 in volleyball players was significantly shorter than that of visual N145 in sedentary subjects. Recent studies have shown that visual N145 and somatosensory P100 waveforms reflect higher-level attentional/cognitive aspects of sensory processing (Tanaka et al. 2008). On the basis of these findings, long-term training may impact not only sensory processing but also neural circuits for perception and cognitive processes necessary for decision making.

Also, the P100 and N140 amplitudes were larger in the reaction condition than in the passive condition in both groups, suggesting that the efferent mechanism might work in both groups to increase the responsiveness of neural population involved in the generation of P100 and N140. Several studies have reported the similar efferent mechanism using EEG and MEG studies (Kida et al. 2006a, b). However, because there was no difference in the P100 and N140 amplitude between two groups, skill-specific training may mainly affect neural circuits but not neural population in the long-latency SEPs.

The neural mechanism underlying decrease in latencies of P100 and N140 should be discussed. The generators for the P100 and N140 have remained unclear. Some MEG (Hari et al. 1993; Mauguiere et al. 1997; Yamashiro et al. 2009) and intracranial studies (Allison et al. 1989; Frot and Mauguiere 1999) have reported activation in secondary somatosensory cortex (SII) range from 70 to 140 ms to the somatosensory stimulation, and this response might be compatible with somatosensory P100. On the other hand, N140 following somatosensory stimulation was mainly generated from anterior cingulate cortex (ACC) in EEG studies (Waberski et al. 2002; Tanaka et al. 2008). Tanaka et al. (2008) suggested that SII (P100) and ACC (N140) activity reflect receiving of sensory information from earlier stage and involuntary shift of attention perceived stimuli, respectively. This finding implies that somatosensory processing has finished by signals arriving to SII. Inui et al. (2004) showed that serial mode of processing both through the postcentral gyrus and through the primary and secondary somatosensory cortices. Since decrease in latency of P100 in baseball group occurred, we speculate that there was opening of shortcut circuits in the earlier stage arriving to SII. Yoshimura et al. (2007) examined the behavior of electrically evoked signals from primary visual cortex to granular retrosplenial cortex in rat brain slices under caffeine application, which resulted in repetitive electrical stimulations opening a new shortcut circuit in a use-dependent manner. Therefore, long-term training may open a new shortcut circuit in human somatosensory cortex.

Here, we present the first evidence for shortening effects in the long-latency SEPs. The results of our study are generally consistent with previous findings for the visual modality. Furthermore, there were significant positive correlations between RT and P100 and N140 latencies. Previous EEG studies (Kekoni et al. 1996; Kida et al. 2004, 2006a; Yamashiro et al. 2008) showed that P100 and N140 reflect passive attention and automatic change detection systems. Therefore, P100 and N140 may be closely related to automatic perception of somatosensory stimulation. In support of this idea, the latencies of P100 and N140 were not significantly different under passive and RT conditions in our study.

In conclusion, the present study reveals that athletes who require finer sensory discrimination and motor skills of the hands and fingers (baseball players) exhibit earlier long-latency SEPs and shorter RTs compared with those who do not require such skills (soccer players, runners, and swimmers). Accordingly, specific and long-term training that engages the cortical somatosensory hand area may alter these neural networks to facilitate motor execution and rapid decision making.

Acknowledgments

This study was supported by a grant-in-aid for young scientists (B) from the Japan Society for the Promotion of Science (JSPS) and a grant-in-aid for Advanced Research from Niigata University of Health and Welfare.

Copyright information

© Springer-Verlag Berlin Heidelberg 2012