Introduction
Influence of Physical Fitness on Human Sensory Long-Term Potentiation
Whether searching popular media or scientific research, the relationship between exercise and cognition is not hard to come by. At the behavioral level, the impact of exercise on cognition has been assessed across a range of cognitive tasks and using a variety of measures for physical fitness, and physical activity (Ferris, Williams, & Shen, Reference Ferris, Williams and Shen2007; Hoffman et al., Reference Hoffman, Blumenthal, Babyak, Smith, Rogers, Doraiswamy and Sherwood2008; Lautenschlager et al., Reference Lautenschlager, Cox, Flicker, Foster, van Bockxmeer, Xiao and Almeida2008; van Uffelen, Chinapaw, van Mechelen, & Hopman-Rock, Reference van Uffelen, Chinapaw, van Mechelen and Hopman-Rock2008; Williamson et al., Reference Williamson, Espeland, Kritchevsky, Newman, King, Pahor and Miller2009; Winter et al., Reference Winter, Breitenstein, Mooren, Voelker, Fobker, Lechtermann and Floel2007). Despite huge variations in the methodology and results of individual studies, the over-arching conclusion from such research has been relatively consistent: increased physical activity is associated with increases in performance on tasks of executive function, processing speed, or memory (Angevaren, Aufdemkampe, Verhaar, Aleman, & Vanhees, Reference Angevaren, Aufdemkampe, Verhaar, Aleman and Vanhees2008; Chang, Labban, Gapin, & Etnier, Reference Chang, Labban, Gapin and Etnier2012; Smith et al., Reference Smith, Blumenthal, Hoffman, Cooper, Strauman, Welsh-Bohmer and Sherwood2010). Be that as it may, there remains much to be elucidated regarding the mechanisms through which physical activity influences cognition, for which a greater understanding of the neurological underpinnings is required.
To address this issue, a growing body of neuroimaging research has begun to assess the changes in neuronal structure and function associated with exercise. MRI studies have demonstrated an increase in cortical grey and white matter volume in response to exercise programs (Chaddock et al., Reference Chaddock, Erickson, Prakash, Kim, Voss, VanPatter and Hillman2010; Erickson et al., Reference Erickson, Prakash, Voss, Chaddock, Hu, Morris and Kramer2009, Reference Erickson, Voss, Prakash, Basak, Szabo, Chaddock and Kramer2011). Additionally, Colcombe et al. (Reference Colcombe, Kramer, Erickson, Scalf, McAuley, Cohen and Elavsky2004) have indicated that high cardiovascular fitness (measured both cross-sectionally and after a 12-month aerobic exercise program) is associated with an increase in BOLD activation within the frontal and parietal lobes, as well as a decreased activation in the anterior cingulate cortex during a modified version of the Erickson Flanker Paradigm. Using a similar paradigm, Liu-Ambrose, Nagamatsu, Voss, Khan, and Handy (Reference Liu-Ambrose, Nagamatsu, Voss, Khan and Handy2012) also identified functional changes in the frontal and temporal lobes in elderly participants following 12-months of resistance-training. Electrophysiological (EEG) studies indicate that regular, intensive physical activity is associated with an increase in baseline theta, alpha, and beta activation (Lardon & Polich, Reference Lardon and Polich1996), as well as an increased amplitude of the P300 component of the visually evoked potential (VEP; Polich & Lardon, Reference Polich and Lardon1997). Increased physical activity is also associated with a decreased latency of this same VEP across a range of tasks (Dustman et al., Reference Dustman, Emmerson, Ruhling, Shearer, Steinhaus, Johnson and Shigeoka1990; Hillman, Belopolsky, Snook, Kramer, & McAuley, Reference Hillman, Belopolsky, Snook, Kramer and McAuley2004; Hillman, Kramer, Belopolsky, & Smith, Reference Hillman, Kramer, Belopolsky and Smith2006). Together, the above body of research indicates that physical activity influences not only the structure, but also the function of the human brain across age groups.
At the cellular level, alterations in synaptic plasticity are understood to facilitate the gross brain changes suggested by the above studies. Specifically, increases in physical activity have been associated with increased neurogenesis in the rodent and human dentate gyrus (Pereira et al., Reference Pereira, Huddleston, Brickman, Sosunov, Hen, McKhann and Small2007; Van Praag, Shubert, Zhao, & Gage, Reference van Praag, Shubert, Zhao and Gage2005). Additionally, the process of long-term potentiation (LTP) has been of increasing interest across the field of physical activity and cognition. LTP refers to an activity-dependent increase in synaptic efficacy following sustained stimulation between neurons (Cooke & Bliss, Reference Cooke and Bliss2006; Teyler, Reference Teyler2000). Typically, assessments of LTP have involved in vitro or in vivo application of high frequency electrical stimulation to the rodent hippocampus, which induces an enhancement of the post-synaptic response from the affected synapses (Bliss & Lømo, Reference Bliss and Lømo1973; Figurov, Pozzo-Miller, Olafsson, Wang, & Lu, Reference Figurov, Pozzo-Miller, Olafsson, Wang and Lu1996; Harris, Ganong, & Cotman, Reference Harris, Ganong and Cotman1984; Teyler & DiScenna, Reference Teyler and DiScenna1987). Conversely, low frequency stimulation leads to an opposing decrease in synaptic efficacy, a process referred to as long-term depression (LTD; Bliss & Cooke, Reference Bliss and Cooke2011; Heynen, Bear, & Abraham, Reference Haynen, Bear and Abraham1995; Mulkey & Malenka, Reference Mulkey and Malenka1992; Xu, Anwyl, & Rowan, Reference Xu, Anwyl and Rowan1997). The above research has also established the presence of Hebbian characteristics in LTP and LTD, such as input specificity, associativity, longevity, saturation, and N-methyl-D-aspartate receptor (NMDAR) dependence (Clapp, Eckert, Teyler, & Abraham, Reference Clapp, Eckert, Teyler and Abraham2006; Cooke & Bliss, Reference Cooke and Bliss2006). In humans, LTP and LTD have only been demonstrated in vitro due to the invasive nature of stimulating hippocampal neurons (Beck, Goussakov, Lie, Helmstaedter, & Elger, Reference Beck, Goussakov, Lie, Helmstaedter and Elger2000; Chen et al., Reference Chen, Lee, Kato, Spencer, Shepherd and Williamson1996). Based on such findings, LTP and LTD represent the primary candidates for modelling the neural basis of learning and memory formation in the mammalian brain (Cooke & Bliss, Reference Cooke and Bliss2006; Teyler & DiScenna, Reference Teyler and DiScenna1987), and are thus of interest in understanding the impact of physical fitness on cognition.
Rodent studies have demonstrated the relationship between increased physical activity (both forced and voluntary) and an increase in memory performance (O’Callaghan, Ohle, & Kelly, Reference O’Callaghan, Ohle and Kelly2007; van Praag, Christie, Sejnowski, & Gage, Reference van Praag, Christie, Sejnowski and Gage1999) as well as Hebbian LTP (Farmer et al., Reference Farmer, Zhao, Van Praag, Wodtke, Gage and Christie2004; O’Callaghan, et al., Reference O’Callaghan, Ohle and Kelly2007; van Praag et al., Reference van Praag, Christie, Sejnowski and Gage1999, Reference van Praag, Shubert, Zhao and Gage2005). However, due to the invasive nature of the above procedures, to date few in vivo human studies have been conducted, limiting our understanding of the relationship between physical activity and LTP in the human brain. However, after Heynen and Bear (Reference Heynen and Bear2001) demonstrated in rodents that visually induced alterations in neural activation could be observed following rapid stimulation, Teyler et al. (Reference Teyler, Hamm, Clapp, Johnson, Corballis and Kirk2005) presented one of the first non-invasive measures of synaptic plasticity in the human visual cortex. Using electroencephalography scalp recordings, Teyler et al. (Reference Teyler, Hamm, Clapp, Johnson, Corballis and Kirk2005) measured VEPs during the presentation of a checkerboard stimulus flashed to the right or left visual hemifield at a low temporal frequency (1 Hz). After being presented with a high-frequency (9 Hz) “tetanic” presentation of the checkerboard stimulus, the amplitude of occipital lobe VEPs increased in both hemispheres. Closer examination revealed that this increase in amplitude was specific to the N1b component of the VEP; a bilateral, negative event-related potential (ERP) component occurring approximately 170 to 190 milliseconds following stimulus onset. The selective increase in the amplitude of the N1b following tetanization suggests an LTP-like alteration of the VEP. This has been replicated in several subsequent human and rodent studies, which have also demonstrated the presence of the Hebbian characteristics of LTP identified in early rodent studies, such as input-specificity (McNair et al., Reference McNair, Clapp, Hamm, Teyler, Corballis and Kirk2006; Ross et al., Reference Ross, McNair, Fairhall, Clapp, Hamm, Teyler and Kirk2008), locality in the cortex (Clapp, Zaehle, et al., Reference Clapp, Zaehle, Lutz, Marcar, Kirk, Hamm and Jancke2005), longevity, NMDAR dependence (Clapp et al., Reference Clapp, Eckert, Teyler and Abraham2006), and a dependence on neurotrophin expression (Lamb et al., Reference Lamb, McKay, Thompson, Hamm, Waldie and Kirk2014; Thompson et al., Reference Thompson, McNair, Kennedy, Antia, Wannenburg, Shelling and Ian2011). Additionally, an equivalent paradigm has demonstrated similar LTP-like alterations in the N1 of auditory evoked potentials (AEPs) following tetanization of auditory tone pips (Clapp, Kirk, Hamm, Shepherd, & Teyler, Reference Clapp, Kirk, Hamm, Shepherd and Teyler2005; Zaehle, Clapp, Hamm, Meyer, & Kirk, Reference Zaehle, Clapp, Hamm, Meyer and Kirk2007) and a recent rodent study has also demonstrated similar long-lasting, NMDAR dependent potentiation following somatosensory stimulation (Han, Huang, Sun, Duan, & Yu, Reference Han, Huang, Sun, Duan and Yu2015). Therefore, while there has been some inconsistency in inducing changes in sensory evoked potentials via sensory tetanization (e.g., Eckert, Guevermont, Williams, & Abraham, Reference Eckert, Guevremont, Williams and Abraham2013), the aforementioned body of human and rodent research does support the potential for inducing LTP-like alterations using non-invasive sensory stimulation (Clapp, Hamm, Kirk, & Teyler, Reference Clapp, Hamm, Kirk and Teyler2012; Kirk et al., Reference Kirk, McNair, Hamm, Clapp, Mathalon, Cavus and Teyler2010). As such, this non-invasive potentiation of sensory ERPs presents an exciting method for investigating the link between exercise and LTP in humans.
Using a similar human LTP paradigm to that first presented by Teyler et al. (Reference Teyler, Hamm, Clapp, Johnson, Corballis and Kirk2005), the aim of the current study was to pilot an assessment of the impact of physical activity on sensory LTP. ERP data were collected and compared between two groups of participants differing in self-reported level of physical activity: a high-activity group, and a low-activity group. It was hypothesized that highly active participants would show greater potentiation of tetanized stimuli than less active participants, as evidenced by greater increases in the amplitude of the N1b both immediately following tetanization, as well as after a 30-min break.
Method
Participants
Twenty-one participants were recruited through advertisement at the University of Auckland, New Zealand. Due to the use of EEG and rapid visual stimulation in the current study, only participants who reported that they did not suffer from any neurological conditions, epilepsy, or migraine were included. All participants had normal or corrected to normal vision and provided informed written consent. BMI and demographic details were collected and are presented in Table 1. All were normally distributed (D=p≥.05). All experimental procedures were approved by the University of Auckland Human Subject Ethics Committee (ref 2011-063).
participant demographics
International Physical Activities Questionnaire
The International Physical Activities Questionnaire (IPAQ)-Long Form was used as a measure of self-reported physical activity. Since its development in 1998, the IPAQ has been used extensively across the globe as a self-report measure of physical activity (Craig et al., Reference Craig, Marshall, Sjöström, Bauman, Booth, Ainsworth and Sallis2003). While there has been some inconsistency (e.g., Maddison et al., Reference Maddison, Mhurchu, Jiang, Vander Hoorn, Rodgers, Lawes and Rush2007), the IPAQ has demonstrated both reliability and validity when compared with accelerometers (Boon, Hamlin, Steel, & Ross, Reference Boon, Hamlin, Steel and Ross2008; Craig et al., Reference Craig, Marshall, Sjöström, Bauman, Booth, Ainsworth and Sallis2003) and measures of maximal oxygen volume consumption (Fogelholm et al., Reference Fogelholm, Malmberg, Suni, Santtila, Kyröläinen, Mäntysaari and Oja2006). It was, therefore, deemed an appropriate tool for self-report physical activity in this current study.
As has been done in previous studies (Fogelholm et al., Reference Fogelholm, Malmberg, Suni, Santtila, Kyröläinen, Mäntysaari and Oja2006), the IPAQ was used to separate participants into two groups based on self-reported levels of daily activity; a “high” activity (11 participants) group and a “low” activity (10 participants) group. Activity levels were calculated in accordance with the IPAQ guidelines for analysis (International Physical Activity Questionnaire, 2005). Briefly, overall physical activity was calculated as MET-min/week. METs (metabolic equivalent of task) are multiples of resting metabolic rates dependent on type of activity and are used to estimate energy requirements for varying tasks (Hagstromer, Oja, & Sjostrom, Reference Hagstromer, Oja and Sjostrom2006). As the IPAQ specifically asks about activity within the past 7 days, participants were also asked whether the past 7 days were an accurate reflection of exercise over the preceding 2 months, to which all participants agreed.
Apparatus
EEG recordings were carried out in an electrically shielded room and measured from a 128-channel Ag/AgCl electrode net (Electrical Geodesics Inc., Eugene, OR) at a continuous 250 Hz sampling rate (0.1–100 Hz analogue bandpass filter). All electrode impedances were below 50 kΩ, which is considered an acceptable level for this system (Ferree, Luu, Russell, & Tucker, Reference Ferree, Luu, Russell and Tucker2001). A common vertex reference (Cz) was used to acquire EEG, which was later referenced to the average off-line.
Stimuli
The visual stimuli were circular horizontally and vertically oriented sine gratings with a one cycle per degree (cpd) spatial frequency. The gratings were presented at full contrast in the center of the screen against a grey background subtending a diameter of 8° of visual angle (size: 9.6×9.6 cm, 272×272 pixels; Figure 1A). Stimuli were presented on a SVGA computer monitor (1024×768 pixel resolution; 60 Hz refresh rate) at a distance of 57 cm. Presentation was controlled using E-Prime v1.1 (Psychology Software Tools).
(a) The visual stimuli were 9.6×9.6 cm circular horizontal and vertical sine gratings (actual size) with a spatial frequency of one cycle per degree and 100 percent contrast presented against a grey background. (b) A schematic representation of the experimental procedure. Participants were presented with two pre-tetanus (baseline) blocks, a photic tetanus block, two early post-tetanus blocks, a 30-min rest period and lastly, two late post-tetanus blocks. The pre-tetanus blocks, early post-tetanus blocks, and late post-tetanus blocks each consisted of 120 presentations of horizontal and vertical circular sine gratings at a frequency of 0.67–1 Hz. The photic tetanus involved 1000 presentation of a single orientation at a frequency of 8.6 Hz. (c) Approximate location of left and right hemisphere electrodes (centered on approximately P7 and P8 under the 10–20 system) used to measure amplitude of N1b component.
Procedure
LTP was assessed via our previously established paradigm using visual-evoked potential EEG scalp-recording (McNair et al., Reference McNair, Clapp, Hamm, Teyler, Corballis and Kirk2006; Ross et al., Reference Ross, McNair, Fairhall, Clapp, Hamm, Teyler and Kirk2008; Teyler et al., Reference Teyler, Hamm, Clapp, Johnson, Corballis and Kirk2005; Thompson et al., Reference Thompson, McNair, Kennedy, Antia, Wannenburg, Shelling and Ian2011). As demonstrated in Figure 1B, the procedure consisted of two pre-tetanus blocks, a photic tetanus, two early post-tetanus blocks, a 30-min rest period, and two late post-tetanus blocks. The pre-tetanus, early post-tetanus, and late post-tetanus blocks each consisted of 120 presentations of each of the horizontal and vertical sine gratings, at a temporal frequency of 0.67–1 Hz (stimulus presentation ~33 ms, jittered ISI of 1000–1500 ms). The photic tetanus consisted of a rapid (8.6 Hz) presentation of 1000 horizontal or vertical sine gratings (counterbalanced between participants; jittered ISI of 67–100 ms). This frequency has been shown to induce the alpha rhythm but not be beyond flicker fusion rate. Following the photic tetanus, participants were asked to close their eyes for 2 min to allow retinal after-images to dissipate. During the 30-min rest period, participants were instructed to sit quietly with their eyes closed.
Analysis
EEG data for each participant were segmented into 600 ms time epochs (100 ms before and 500 ms after stimulus onset), bandpass filtered offline [0.10–30 Hz, bi-directional three-pole Butterworth filter (Alarcon, Guy, & Binnie, Reference Alarcon, Guy and Binnie2000)], and epochs were averaged by condition according to orientation (horizontal or vertical), and block (pre-tetanus, early post-tetanus, late post-tetanus). Epochs were baselined to the 100 ms pre-stimulus period. Epochs with eye blinks or artifacts were rejected (range: 0–10% of epochs per participant, no significant differences between groups). The amplitude of the N1b component of the pre-tetanus, early post-tetanus and late post-tetanus VEPs were averaged for clusters of electrodes positioned over the left and right visual cortex (Figure 1C). In accordance with previously validated measures (McNair et al., Reference McNair, Clapp, Hamm, Teyler, Corballis and Kirk2006), the amplitude of the N1b component was defined as the mean amplitude of the section of the evoked potential extending from the peak of the N1 component to halfway between the N1 and P2 components. For each participant individually, the N1b was identified (latencies were between 100 ms and 200 ms for all participants) and the magnitude of LTP was calculated by subtracting pre-tetanus N1b amplitude from early and late post-tetanus amplitudes. There were no significant differences between pre- and post-tetanus amplitudes for the non-tetanized gratings (p≥.05) and, therefore, these were excluded from further analyses.
Changes in amplitude as a function of level of activity were analyzed using a 2×2 Split-Plot analysis of variance (ANOVA) with Activity Level (low or high) as the between-subjects variable and Time (early or late post tetanus) as the within-subjects variable (alpha level p≤.05). All statistical analyses were performed using the statistical analysis software package SPSS 20.0. A Greenhouse-Geisser correction was used to control for violations of sphericity where required.
Results
There were no significant differences in BMI between the two groups, t(19)=−0.83, p=.417. Conversely, as expected, the IPAQ scores for the Low Activity group were significantly lower than the High Activity group, t(11.54)=−5.65, p<.01. Figure 2 presents the difference in N1b amplitude between pre-tetanus and the early and late post-tetanus blocks for the two groups separately. The two-way split-plot ANOVA revealed a significant main effect for Time, indicating that LTP amplitude was greater for the early post-tetanus block (M=−0.70 µV; SE=.17 µV) than the late post-tetanus block (M=−.39 µV; SE=0.15 µV), F(1,19)=7.02, p=.016, ηp 2=.270. Importantly, the interaction between Activity Level and Time was significant, F(1,19)=10.705, p=.004, ηp 2=.360. For the early post-tetanus block, there was no significant difference in N1b amplitude between the High Activity (M=−0.77 µV; SE=0.24 µV) and the Low Activity group (M=−0.64 µV; SE=0.25 µV; p>.05). However, while the N1b remained potentiated (negative compared to pre-tetanus baseline) in the late post-tetanus block for the High Activity group (M=−0.84; SE=0.21), it returned to baseline for the Low Activity group (M=0.07; SE=0.22; p=.008), leading to significant group differences in the late post-tetanus block. The main effect for Activity Level was non-significant, F(1,19)=2.96, p=.102, ηp 2=.134. These trends are illustrated for one subject in each group in Figure 3.
Bar graph showing the mean amplitude of N1b for early and late LTP post-tetanus for the high and low activity groups (error bars depicting standard error). There was a significant difference between early and late LTP for the low activity group. There was a significant difference between the low and high activity groups for late post-tetanus LTP. * indicates a statistically significant effect (p≤.05).
Exemplars from single subjects showing the trends described in Figure 2. Negative plotted up. (a, c) ERPs from pre-tetanus (thin black line), early (thick gray line), and late (thick black line) post-tetanus for the high and low activity groups, respectively. (b, d) topographic maps at the mid N1b of the pre-tetanus potential for the high and low activity groups, respectively. Time on x-axis in ms. Scale bar at 0 ms ±1 μV.
Discussion
The aim of the current study was to use an established non-invasive measure of LTP to pilot an assessment of the impact of physical activity on neuroplasticity. Using the IPAQ, participants were divided into two groups (highly active and less active) and ERPs were recorded during the presentation of “tetanized” and “non-tetanized” sine gratings. Both groups demonstrated a significant “early” LTP effect. That is, the tetanus resulted in an increase in the amplitude of the N1b component of the VEPs immediately following tetanization. Importantly, however, there was a significant difference in “late” LTP (measured 30 min after induction) between the High Activity group and the Low Activity group. The High Activity group demonstrated significantly greater potentiation of the N1b in the late post-tetanus block than the Low Activity group. This study, therefore, presents the first in vivo evidence to suggest that chronic levels of physical activity can influence the maintenance of LTP-like changes in the human visual cortex, and thus illuminates potential neurological mechanisms underlying the relationship between exercise and cognition.
The overall increase in amplitude of the N1b following tetanization for both groups supports previous research indicating that an LTP-like response can be invoked through non-invasive visual stimulation (Clapp, Zaehle, et al., Reference Clapp, Zaehle, Lutz, Marcar, Kirk, Hamm and Jancke2005; Clapp, Kirk, et al., Reference Clapp, Kirk, Hamm, Shepherd and Teyler2005; McNair et al., Reference McNair, Clapp, Hamm, Teyler, Corballis and Kirk2006; Ross et al., Reference Ross, McNair, Fairhall, Clapp, Hamm, Teyler and Kirk2008; Zaehle et al., Reference Zaehle, Clapp, Hamm, Meyer and Kirk2007). Previous assessments of LTP in rodents have largely taken place within the hippocampus, a structure undoubtedly involved in memory (e.g., Bliss & Lømo, Reference Bliss and Lømo1973; Figurov et al., Reference Figurov, Pozzo-Miller, Olafsson, Wang and Lu1996; Harris et al., Reference Harris, Ganong and Cotman1984). Conversely, the current study focused on the sensory cortex which allowed the implementation of a non-invasive design. There is ample evidence indicating that neuroplasticity in the visual cortex represents the same underlying processes as hippocampal LTP (for review, Cooke & Bear, Reference Cooke and Bear2014). Additionally, both rodent and human studies using versions of the current paradigm have demonstrated Hebbian characteristics in these LTP-like changes such as specificity (McNair et al., Reference McNair, Clapp, Hamm, Teyler, Corballis and Kirk2006; Ross et al., Reference Ross, McNair, Fairhall, Clapp, Hamm, Teyler and Kirk2008) longevity (Clapp et al., Reference Clapp, Eckert, Teyler and Abraham2006), and NMDA dependence (Clapp et al., Reference Clapp, Eckert, Teyler and Abraham2006). However, it must be acknowledged that future studies are required to elucidate the generalizability of these LTP-like changes to the rest of the cortex. As such, while potentiation of VEPs may not directly represent declarative memory formation, this does represent activity-dependent neuroplasticity in the human cortex, and thus has implications in our understanding of the neural basis of cognition and memory. Importantly, this provides direction for potential intervention studies in the future.
This study, therefore, presents an insight into the potential neuronal mechanisms behind the large body of human and rodent research indicating that physical activity increases cognitive performance (e.g., Colcombe et al., Reference Colcombe, Kramer, Erickson, Scalf, McAuley, Cohen and Elavsky2004; Ferris et al., Reference Ferris, Williams and Shen2007; Lautenschlager et al., Reference Lautenschlager, Cox, Flicker, Foster, van Bockxmeer, Xiao and Almeida2008; Winter et al., Reference Winter, Breitenstein, Mooren, Voelker, Fobker, Lechtermann and Floel2007). As expected, the highly active group demonstrated a greater increase in the amplitude of the N1b following tetanization than the less active group; thus indicating greater LTP-like changes in the VEP. As LTP is believed to underlie memory formation and maintenance (Cooke & Bliss, Reference Cooke and Bliss2006; Teyler & DiScenna, Reference Teyler and DiScenna1987), this result suggests that greater levels of physical activity increase neural plasticity, which may in turn increase an individual’s capacity for learning and memory. However, an unexpected finding in the current study was that this group difference was only significant for the late post-tetanus block, 30 min after tetanization. This suggests that while there is no influence of chronic physical activity on the induction of experience-dependent plasticity, higher levels of physical activity may be beneficial for the maintenance of these changes. While it is unclear why differences in activity level specifically targeted late LTP, several potential hypotheses can be proposed. Largely, these hypotheses are centered on the potential mediating effect of brain derived neurotrophic factor (BDNF) in exercise-induced cognitive improvements and LTP.
BDNF is one of four neurotrophins found in the mammalian central nervous system that are responsible for the regulation of neuronal growth, maintenance, and survival, as well as sustaining the malleable state required for learning and memory (Reichardt, Reference Reichardt2006; Tyler, Alonso, Bramham, & Pozzo-Miller, Reference Tyler, Alonso, Bramham and Pozzo-Miller2002). Importantly, the activity-dependent nature of BDNF renders it a primary candidate for modulating the induction and maintenance of LTP (Bekinschtein et al., Reference Bekinschtein, Cammarota, Katche, Slipczuk, Rossato, Goldin and Medina2008; Poo, Reference Poo2001; Tyler et al., Reference Tyler, Alonso, Bramham and Pozzo-Miller2002), and it is believed to be involved in the regulation of pre-synaptic neurotransmitter release, as well as post-synaptic protein synthesis and transcription (Bekinshtein, Cammarota, & Medina, Reference Bekinschtein, Cammarota and Medina2014; Christie et al., Reference Christie, Brennan, Kannangara, Robillard, Shin and Titterness2008). Rodent models have illustrated a positive relationship between physical fitness and expression of hippocampal BDNF (Cotman & Berchtold, Reference Cotman and Berchtold2002; Garza, Ha, Garcia, Chen, & Russo-Neustadt, Reference Garza, Ha, Garcia, Chen and Russo-Neustadt2004; Gómez-Pinilla, Ying, Roy, Molteni, & Edgerton, Reference Gómez-Pinilla, Ying, Roy, Molteni and Edgerton2002; Marais, Stein, & Daniels, Reference Marais, Stein and Daniels2009; Neeper, Gómez-Pinilla, Choi, & Cotman, Reference Neeper, Gómez-Pinilla, Choi and Cotman1996). A limited number of human studies have also exemplified the link between both chronic and acute increases in physical activity and increased BDNF expression (Erickson et al., Reference Erickson, Voss, Prakash, Basak, Szabo, Chaddock and Kramer2011; Ferris et al., Reference Ferris, Williams and Shen2007; Zoladz et al., Reference Zoladz, Pilc, Majerczak, Grandys, Zapart-Bukowska and Duda2008), as well as an increase in visually evoked LTP in individuals expressing greater levels of BDNF (based on the BDNF Val66Met polymorphism; Thompson et al., Reference Thompson, McNair, Kennedy, Antia, Wannenburg, Shelling and Ian2011, 2015; Lamb et al., Reference Lamb, McKay, Thompson, Hamm, Waldie and Kirk2014).
One possible mechanism by which BDNF may be pivotal in the current results is that activity-induced increases in BDNF expression may specifically impact LTP maintenance (synthesis and transcription), while LTP induction remains unaffected. This would only be apparent in the late post-tetanus block of the current study where LTP maintenance is required. However, as previous human research has indicated a general increase in BDNF expression with increased exercise, it is unclear why this would be the case (Ferris et al., Reference Ferris, Williams and Shen2007; Zoladz et al., Reference Zoladz, Pilc, Majerczak, Grandys, Zapart-Bukowska and Duda2008). As an alternative hypothesis, in vitro human studies have indicated that low frequency stimulation (1 Hz) induces LTD (Beck et al., Reference Beck, Goussakov, Lie, Helmstaedter and Elger2000; Chen et al., Reference Chen, Lee, Kato, Spencer, Shepherd and Williamson1996) and, as proposed by Teyler et al. (Reference Teyler, Hamm, Clapp, Johnson, Corballis and Kirk2005) and Ross et al. (Reference Ross, McNair, Fairhall, Clapp, Hamm, Teyler and Kirk2008), it is possible that the baseline stimulation used in the current paradigm (0.67–1 Hz) induces LTD, thus actively depotentiating (reducing) the previously enhanced N1b. Importantly, Ikegaya, Ishizaka, and Matsuki (Reference Ikegaya, Ishizaka and Matsuki2002) demonstrated a significant reduction in the LTD induced by low frequency stimulation in hippocampal slices bathed in BDNF. This indicates that increased BDNF buffers against active depotentiation, and an increased expression of BDNF in highly active individuals may facilitate this protective effect. It must be noted that the current pilot study cannot delineate between the above proposals. However, it lays the foundations for future studies incorporating neurotrophin expression and gene interactions such as the BDNF Val66Met polymorphism into our understanding of the cellular and molecular mediators of exercise-induced LTP.
It is important to note that, while the N1b in the late post-tetanus block for the low activity group appears to return to an amplitude almost identical to that recorded pre-tetanus, this does not indicate that low fitness is associated with a complete inability to maintain LTP. As suggested above, individuals from the low activity group may be more susceptible to active depotentiation induced by low frequency stimulation. Additionally, Farmer et al. (2004) indicated that increased fitness level lowers the threshold for the induction of long-lasting LTP. Therefore, highly active participants may be more susceptible to LTP following a 8.6 Hz tetanus than less active participants. As such, the current results are not inconsistent with previous rodent studies demonstrating an enhancement of LTP magnitude with increased physical fitness (O’Callaghan et al., Reference O’Callaghan, Ohle and Kelly2007; van Pragg et al., 1999) as these results indicate an enhancement in the maintenance of experience-dependent plasticity for highly active individuals. It is anticipated that future research will further illuminate the mechanisms underlying these differences.
The current pilot study presents the first evidence for fitness related differences in the maintenance of LTP-like changes in the human visual cortex. However, certain caveats must be highlighted when interpreting the current results. First, it is important to acknowledge that the measure of physical activity used in the current study was a self-report measure and not a direct physiological measure of physical fitness. While the IPAQ has shown test–retest reliability and validity as a measure for activity status (Brown, Trost, Bauman, Mummery, & Owen, Reference Brown, Trost, Bauman, Mummery and Owen2004; Craig et al., Reference Craig, Marshall, Sjöström, Bauman, Booth, Ainsworth and Sallis2003), the gold standard for physical fitness is the maximal volume of oxygen consumption (VO2 max) (Garber et al., Reference Garber, Blissmer, Deschenes, Franklin, Lamonte, Lee and Swain2011). Nevertheless, in a comparison of the validity of the IPAQ and VO2 max by ergometer, Fogelholm et al. (Reference Fogelholm, Malmberg, Suni, Santtila, Kyröläinen, Mäntysaari and Oja2006) demonstrated that responses categorized as “high” and “low” physical fitness based on the IPAQ correlated with higher and lower cardiorespiratory fitness (with the exception of a small proportion of young men who over-reported physical activity). Furthermore, the IPAQ has been successfully used to distinguish between physically active and sedentary individuals to measure cognition as a function of fitness level (Kamijo, O’leary, Pontifex, Themanson, & Hillman, Reference Kamijo, O’leary, Pontifex, Themanson and Hillman2010). It would be interesting to explore additional measures of fitness in future research to further elucidate the relationship between fitness and LTP.
Second, previous research has indicated that different types of exercise can have substantially different effects on cognition and neuroplasticity. For example, in a series of experiments, Erickson and colleagues demonstrated a positive relationship between an aerobic exercise program and gray and white matter volume (Colcombe et al., Reference Colcombe, Erickson, Scalf, Kim, Prakash, McAuley and Kramer2006; Erickson et al., Reference Erickson, Prakash, Voss, Chaddock, Hu, Morris and Kramer2009, Reference Erickson, Raji, Lopez, Becker, Rosano, Newman and Kuller2010, Reference Erickson, Voss, Prakash, Basak, Szabo, Chaddock and Kramer2011), while Liu-Ambrose et al. (Reference Liu-Ambrose, Nagamatsu, Graf, Beattie, Ashe and Handy2010) identified a decrease in whole brain volume following a 12-month resistance training intervention. Colcombe and Kramer (Reference Colcombe and Kramer2003) suggested aerobic exercise combined with resistance training may provide the greatest benefit for cognitive function. However, the current study did not differentiate between different types of activity.
Finally, it is important to note the lack of significant difference in BMI between the two participant groups. While several previous studies have demonstrated a relationship between BMI and physical fitness (Frändin & Grimby, Reference Frändin and Grimby1994; Leyk et al., Reference Leyk, Rohde, Gorges, Ridder, Wunderlich, Dinklage and Essfeld2006), this has not been consistent across all studies (Mak et al., Reference Mak, Ho, Lo, Thomas, McManus, Day and Lam2010). Additionally, there are several lifestyle (Tavani, Negri, & La Vecchia, Reference Tavani, Negri and La Vecchia1994), genetic (Speliotes et al., Reference Speliotes, Willer, Berndt, Monda, Thorleifsson, Jackson and Loos2010), and demographic (Leyk et al., Reference Leyk, Rohde, Gorges, Ridder, Wunderlich, Dinklage and Essfeld2006; Tavani et al., Reference Tavani, Negri and La Vecchia1994) factors that can influence BMI beyond physical activity that would be difficult to exclude as confounds if group differences in BMI were to emerge. As highlighted by Leyk et al. (Reference Leyk, Rohde, Gorges, Ridder, Wunderlich, Dinklage and Essfeld2006), while fitness and BMI may correlate, the concepts of being “overweight” and being “physically unfit” are not synonymous (as is true for being “underweight” and being “physically fit”); therefore, one should not be used as a marker or criteria for the other. Be that as it may, future research may wish to investigate the influence of BMI on LTP, as well as their potential interaction.
The non-invasive paradigm created by Teyler et al. (Reference Teyler, Hamm, Clapp, Johnson, Corballis and Kirk2005) has been used to demonstrate that LTP impairments are associated with a range of neuropsychological disorders such as depression (Holderbach, Clark, Moreau, Bischofberger, & Normann, Reference Holderbach, Clark, Moreau, Bischofberger and Normann2007; Normann et al., Reference Normann, Schmitz, Fürmaier, Döing and Bach2007), bipolar-II disorder (Elvsåshagen et al., Reference Elvsåshagen, Moberget, Bøen, Boye, Englin, Pedersen and Andersson2012), and schizophrenia (Mears & Spencer, Reference Mears and Spencer2012). As suggested by Kirk et al. (Reference Kirk, McNair, Hamm, Clapp, Mathalon, Cavus and Teyler2010), numerous neurocognitive disorders may manifest synaptic deficiencies in varying affected cortical areas, and Krystal et al. (Reference Krystal, Tolin, Sanacora, Castner, Williams, Aikins and D’Souza2009) advocate for increased research focusing on psychiatric treatments that could manipulate neuroplasticity in conjunction with behavioral interventions. The results of the current study indicate a potential candidate for the neural basis for exercise-induced cognitive enhancements, and as such, it is worth considering exercise as a non-pharmaceutical option to alterations in neural plasticity.
Acknowledgements
This work was supported by a NZ Royal Society Marsden Grant (I.J.K., grant number 06-UoA-077). There are no conflicts of interest to report.
