Abstract

The ability to discriminate among goal-relevant stimuli tends to diminish when detections must be made continuously over time. Previously, we reported that intensive training in shamatha (focused-attention) meditation can improve perceptual discrimination of difficult-to-detect visual stimuli [MacLean, K. A., Ferrer, E., Aichele, S. R., Bridwell, D. A., Zanesco, A. P., Jacobs, T. L., et al. Intensive meditation training improves perceptual discrimination and sustained attention. Psychological Science, 21, 829–839, 2010]. Here we extend these findings to examine how discrimination difficulty and meditation training interact to modulate event-related potentials of attention and perceptual processing during vigilance. Training and wait-list participants completed a continuous performance task at the beginning, middle, and end of two 3-month meditation interventions. In the first intervention (Retreat 1), the continuous performance task target was adjusted across assessments to match training-related changes in participants' perceptual capacity. In the second intervention (Retreat 2), the target was held constant across training, irrespective of changes in discrimination capacity. No training effects were observed in Retreat 1, whereas Retreat 2 was associated with changes in the onset of early sensory signals and an attenuation of within-task decrements at early latencies. In addition, changes at later stimulus processing stages were directly correlated with improvements in perceptual threshold across the second intervention. Overall, these findings demonstrate that improvements in perceptual discrimination can modulate electrophysiological markers of perceptual processing and attentional control during sustained attention, but likely only under conditions where an individual's discrimination capacity is allowed to exceed the demand imposed by the difficulty of a visual target. These results contribute to basic understanding of the dependence of perceptual processing and attentional control to contextual demands and their susceptibility to directed mental training.

INTRODUCTION

Voluntary attention is limited, and the requirement to sustain attention over time can lead to a substantial decline in performance, especially when attentional demands are high. Indeed, recognition of attentional failings, and the importance of strengthening attention to promote adaptive self-control and psychological well-being (Wadlinger & Isaacowitz, 2011; Kaplan & Berman, 2010) have long informed the development of systems of meditation-based mental training (Wallace, 1999). Shamatha [calm abiding] practices comprise one important class of meditation techniques that are thought to increase practitioners' capacity to sustain attention and to enhance the perceived detail of attended sensations (Dahl, Lutz, & Davidson, 2015; Wallace, 2006). In an initial test of these claims, we previously reported that intensive, full-time training in shamatha meditation improved visual discrimination, leading to apparent increases in successful target detection across minutes of sustained visual attention (MacLean et al., 2010). Since that time, we have accrued additional behavioral evidence that intensive meditation practice may afford benefits for sustained attentional performance and cognitive control (Zanesco, King, MacLean, & Saron, 2013, 2018).

The prospect that sustained attention can be systematically trained holds potentially broad implications for theories of attention and the plasticity of associated cognitive systems (Slagter, Davidson, & Lutz, 2011). However, the neural changes that accompany meditation-related improvements in visual discrimination and vigilance have not been characterized. Event-related potential (ERP) studies, in particular, offer an important complement to behavioral approaches, because neurophysiological information can aid in characterizing stimulus processing stages that likely contribute to reported improvements in performance. With this in mind, we sought to examine how training-related changes in visual discrimination modulate neuroelectric responses underlying bottom–up sensory and top–down attentional influences on accurate stimulus detection during sustained performance.

Investigations of stimulus-evoked brain potentials have demonstrated that attention can modulate a succession of visual ERP components, hypothesized to represent multiple stages of information processing. At early stages, attentional focus can enhance visual sensory signals within 90 msec following stimulus presentation (Baumgartner, Graulty, Hillyard, & Pitts, 2018). Attention can also modulate subsequent early stages such as the visual N1, which is commonly observed as an occipital-parietal negativity occurring from 150 to 200 msec following stimulus onset. The N1 is thought to reflect an early top–down discrimination process, as ERP amplitude at this latency can differentiate attended stimuli on the basis of task relevance (Fedota, McDonald, Roberts, & Parasuraman, 2012; Hopf, Vogel, Woodman, Heinze, & Luck, 2002; Vogel & Luck, 2000). Differential engagement of top–down attentional demands can also influence visual discrimination processes at the latency of N1. For example, N1 amplitude is increased when the rotation difference between serially presented target and nontarget Gabor patches is made difficult to distinguish (Fedota et al., 2012), suggesting that attentional modulation of neural activity at this latency may reflect the operation of sensory gain control mechanisms critical for stimulus discrimination and subsequent processing: As discrimination difficulty increases, early sensory stages engage greater processing resources to discriminate ever smaller feature differences between targets.

At later stages of stimulus processing, attention operates to support goal-related behaviors by maintaining stimulus representations for further evaluation and categorization. These decision-related processes coincide with the P3 (Polich, 2007), a positive scalp voltage deflection occurring at a latency of 300–600 msec. The P3 is enhanced following presentation of attended or correctly identified target stimuli, relative to nontargets, and is therefore thought to reflect later stimulus detection and evaluation processes. Task manipulations that increase cognitive demand, discrimination difficulty, and decision uncertainty prompt a reduction in P3 amplitude (Polich, 2007), supporting the view that the P3 indexes the allocation of limited attentional resources for successful target detection (Kok, 2001).

Attentional and perceptual demands can exert additional processing costs when performance is to be maintained over time. Most prominently, behavioral measures of target detection are known to evince a reliable vigilance decrement or monotonic decline over sustained task performance when stimulus discriminations are made perceptually challenging (Nuechterlein, Parasuraman, & Jiang, 1983) or when working memory load is high (Parasuraman, 1979). Surprisingly little research, however, has addressed the contributions of perceptual processing to event-related correlates of visual stimuli in the context of tasks requiring sustained attention. The available evidence appears to suggest that effects of lapsing vigilance begin with degradation of early discrimination processes, which exert a cascading influence on the quality of subsequent stimulus representations at later processing stages (e.g., Haubert et al., 2018; Boksem, Meijman, & Lorist, 2005; Parasuraman, Warm, & See, 1998). Critically, though, it is unclear how or whether changes in visual discrimination capacity serve to moderate within-task reductions in amplitude of the visual-evoked N1 and latter P3.

Cross-sectional studies have reported faster latency and greater magnitude of early sensory-evoked potentials, including the P1 and N1 responses, in experienced meditation practitioners as compared with meditation-naive individuals (Atchley et al., 2016; van Leeuwen, Singer, & Melloni, 2012). These and further studies have also reported increased P3 amplitude in meditation practitioners relative to naive controls when correctly detecting visual target stimuli (Jo, Schmidt, Inacker, Markowiak, & Hinterberger, 2016; Delgado-Pastor, Perakakis, Subramanya, Telles, & Vila, 2013) and reductions in P3 amplitude to distracting auditory stimuli during periods of meditation (Cahn & Polich, 2009). Together, these findings support the idea that attentional benefits of meditation training may accrue, in part, through facilitation of early sensory gain control mechanisms, buttressing practitioners' ability to efficiently allocate processing resources to attended stimuli, as indexed by modulations of P3 amplitude. But some evidence suggests meditation training may influence alternative attentional mechanisms. Lutz et al. (2009), for example, observed no changes in early sensory responses or P3 amplitude to auditory target stimuli following 3 months of full-time meditation training, although they did provide evidence for increased oscillatory entrainment to sensory input as a mechanism for improved sustained attention in practitioners.

In this study, we examined longitudinal changes in scalp-recorded visual ERPs during 32 min of sustained visual discrimination. Participants were experienced meditators randomly assigned to receive 3 months of meditation training first (Retreat 1) or to serve as wait-list controls and receive training second (Retreat 2). Training was conducted in an intensive residential retreat setting designed to facilitate full-time, uninterrupted meditation practice (King, Conklin, Zanesco, & Saron, 2019). Sustained attention was assessed through a continuous performance task (CPT; MacLean et al., 2009, 2010) shown to induce reliable declines in individuals' ability to discriminate difficult-to-detect rare target stimuli from frequent nontarget stimuli over time. Critically, participant's discrimination capacity was measured before CPT performance and used to manipulate target discriminability on an individual basis.

Behavioral findings for this intervention were described previously by MacLean et al. (2010), in which our research group reported robust improvements in perceptual threshold following both 3-month-long retreats. Yet, we observed no changes in accuracy or vigilance in the CPT when target discriminability was adjusted dynamically across assessments to match participants' perceptual capacity in Retreat 1. Based on this observation, we hypothesized that increased perceptual threshold impaired training participants' ability to improve on the CPT because they were making harder discriminations relative to controls. To examine this possibility directly, we held target stimulus length constant across assessments in Retreat 2 and subsequently observed improvements in discrimination accuracy and an attenuation of the vigilance decrement in the CPT following meditation training. These findings imply that bottom–up perceptual capacities are honed through focused-attention meditation, and the interaction of external task demands and perceptual capacities appears to constrain the deployment of attentional resources over time.

The current study sought to examine the contribution of these moderating influences to the event-related stimulus processing stream. We employed reference-independent analyses of global field power (GFP) and event-related electric field topography (Murray, Brunet, & Michel, 2008) to characterize modulations in ERPs involved in distinguishing correctly discriminated frequent targets from infrequent nontargets. We aimed to clarify whether meditation-related improvements in perceptual discrimination across 3 months of training reflect neuroplastic changes in bottom–up sensory or top–down attentional systems and if such observed effects contribute to changes in processing during the CPT at early sensory or later processing stages. Furthermore, their dependence on contextual perceptual demands and direct association with individuals' perceptual abilities may help elucidate the mechanisms involved in meditation training-related attentional improvements.

METHODS

Participants

Sixty experienced meditation practitioners were recruited and assigned to either an initial training group (n = 30) or a wait-list control group (n = 30) through stratified random assignment. Groups were matched on age (M = 48 years, range = 22–69), sex (28 men, 32 women), handedness (57 right-handed, 3 left-handed), number of prior meditation retreats (M = 14), mean minutes of daily meditation practice (M = 55), and estimates of lifetime experience (M = 2610 total hours; see MacLean et al., 2010, for full recruitment and matching criteria). All participants had normal or corrected-to-normal vision (confirmed using a T2a vision screener from Titmus Optical). During an initial 3-month intervention (Retreat 1), training participants lived and practiced meditation on site at Shambhala Mountain Center (SMC) in Red Feather Lakes, CO. Approximately 3 months later, wait-list controls underwent formally identical training during a second 3-month intervention (Retreat 2; n = 291) at SMC. All study procedures were approved by the institutional review board of the University of California, Davis. All participants gave full informed consent and were compensated $20 per hour of data collection.

Meditation Training

Participants received meditation training and instruction from B. Alan Wallace, a Buddhist teacher, scholar, and contemplative practitioner.2 Training included two styles of Buddhist meditation practice (Wallace, 2006): shamatha techniques that foster calm sustained attention on a chosen object and complementary techniques that generate benevolent aspirations and feelings toward oneself and others (e.g., compassion). Shamatha training involved three primary practices: (1) “mindfulness of breathing,” in which attention is drawn to the tactile sensations of the breath; (2) attending to the arising of mental content (e.g., thoughts, perceptions, sensations), a technique known as “settling the mind into its natural state”; and (3) focusing attention on the sense of awareness itself, known as “shamatha without a sign” (Wallace, 2006). Participants were also encouraged to maintain mindful awareness of their actions and surroundings throughout the day; met twice daily for group practice and discussion; and devoted an average of 6 hr (SD = 1.5) of their remaining daily time to formal solitary shamatha meditation.

Procedure

Initial training participants were assessed at the beginning (preassessment), middle (midassessment), and end (postassessment) of Retreat 1. Wait-list participants were assessed at the beginning, middle, and end of both retreats, first as controls and then as active training participants in Retreat 2. Participants completed a battery of neurocognitive tasks at each assessment in two sound-attenuated, darkened testing chambers located in the building where training participants lived and meditated. During Retreat 1, control participants traveled to SMC 3 days (range = 65–75 hr) before each assessment to acclimatize to the altitude and retreat environment. Between assessments, wait-list controls returned to their ordinary routines at home; when at SMC, they enjoyed the natural environment (e.g., hiking) and were free to communicate with fellow participants but were instructed not to practice meditation intensively.

Perceptual Threshold Procedure

At each assessment, participants first completed a discrimination threshold procedure (∼10 min) designed to calibrate target stimulus line length according to an individual's perceptual discrimination threshold (as in MacLean et al., 2009, 2010). Figure 1 depicts the stimuli and timing for this thresholding procedure and for the full 32-min CPT. Fixation was maintained on a small dot at the center of the screen, where single gray vertical lines appeared one at a time against a black background. Each line stimulus was presented for 150 msec with a 1550–2150 msec variable ISI. Frequent long line nontargets (4.82° visual angle) occurred on 70% of trials, and rare short line targets occurred on 30% of trials. Participants were instructed to respond as quickly and accurately as possible with the left mouse button (right index finger) to rare short line targets and to not respond to frequent long line nontargets. Sound feedback was used to indicate correct and incorrect responses on each trial. A visual mask was presented for the duration of each ISI, composed of many short lines ranging in height (0.28°–0.45° visual angle) throughout a 5.0° × 1.0° space surrounding the fixation dot. A unique mask pattern was generated for each stimulus presentation.

Figure 1. 

Stimuli and timing for the threshold procedure and CPT. Single lines (light gray, 40.29 cd/m2) were presented at the center of the screen against a black background (0.35 cd/m2) while participants fixated on a small yellow dot (shown in white) from a viewing distance of 57 cm. In the threshold procedure, long nontarget lines (4.82°) were presented 70% of the time, and short target lines (range = 2.76°–4.78°) were presented 30% of the time. In the CPT, target frequency was reduced to 10% of stimuli and target line length was determined based on accuracy during the threshold procedure. Each stimulus was presented for 150 msec, and a mask was presented during the variable ISI of 1550–2150 msec. Instructions emphasized quickly and accurately responding to short target lines.

Figure 1. 

Stimuli and timing for the threshold procedure and CPT. Single lines (light gray, 40.29 cd/m2) were presented at the center of the screen against a black background (0.35 cd/m2) while participants fixated on a small yellow dot (shown in white) from a viewing distance of 57 cm. In the threshold procedure, long nontarget lines (4.82°) were presented 70% of the time, and short target lines (range = 2.76°–4.78°) were presented 30% of the time. In the CPT, target frequency was reduced to 10% of stimuli and target line length was determined based on accuracy during the threshold procedure. Each stimulus was presented for 150 msec, and a mask was presented during the variable ISI of 1550–2150 msec. Instructions emphasized quickly and accurately responding to short target lines.

Parameter Estimation through Sequential Testing (PEST; Taylor & Creelman, 1967) was used to estimate the short line target length that could be correctly discriminated at a predetermined accuracy for each participant, defined as perceptual discrimination threshold and reported in units of visual angle (possible range of 2.76°–4.78°).The threshold procedure continued until converging on 85% target accuracy for Retreat 1 preassessment, and 75% target accuracy for all remaining assessments. This change (from 85% to 75% PEST accuracy) was implemented to ensure high task demand after some participants unexpectedly failed to show within-task decrements in performance at the initial assessment (see MacLean et al., 2010).

Continuous Performance Task

The 32-min CPT was completed immediately following the perceptual threshold procedure. Participants were instructed to respond to rare short line targets (10% of trials; 96 total target trials) and not respond to frequent long line nontargets (90% of trials; 864 total nontarget trials). Stimulus and response parameters were identical to the perceptual threshold procedure, except that target line length remained constant throughout the task, and sound feedback was not present. “Hits” were defined as correct responses to rare targets, and “correct rejections” as correct nonresponses to nontargets.

Two different target length setting manipulations were employed across retreats. During Retreat 1, CPT target length was individually determined (reparameterized) at each assessment based on participants' measured PEST discrimination threshold. In Retreat 2, target length was only reparameterized at preassessment and then held constant to this value across the remaining assessments: Although discrimination threshold was measured at all Retreat 2 assessments, it was only determinative of target length at the onset of training. See Table 1 for a summary of this procedure.

Table 1. 
CPT Task Parameters and Descriptive Statistics for the Discrimination Threshold Procedure
 Retreat 1Retreat 2
PreMidPostPreMidPost
Reparameterized Yes Yes Yes Yes No No 
PEST difficulty 85% 75% 75% 75% 75% 75% 
Training threshold 3.760 (0.271) 4.186 (0.197) 4.273 (0.181) – – – 
Wait-list threshold 3.782 (0.239) 4.092 (0.351) 4.118 (0.239) 4.108 (0.273) 4.239 (0.165) 4.254 (0.162) 
 Retreat 1Retreat 2
PreMidPostPreMidPost
Reparameterized Yes Yes Yes Yes No No 
PEST difficulty 85% 75% 75% 75% 75% 75% 
Training threshold 3.760 (0.271) 4.186 (0.197) 4.273 (0.181) – – – 
Wait-list threshold 3.782 (0.239) 4.092 (0.351) 4.118 (0.239) 4.108 (0.273) 4.239 (0.165) 4.254 (0.162) 

Means and standard deviations (in parentheses) for discrimination threshold for the Retreat 1 training group (n = 26) and the wait-list control group (n = 25 for Retreat 1; n = 27 for Retreat 2). Reparameterized indicates whether the CPT target line length was determined based the threshold procedure at that same assessment (Yes) or whether it was preset (No) to a prior threshold. PEST difficulty refers to the % target accuracy used to determine the short line target length for each participant. Discrimination threshold (training and wait-list control) is the achieved mean visual angle determined by the PEST procedure.

EEG Data Collection and Processing

EEG was recorded at 2048 Hz from 88 electrodes (equidistant montage, www.easycap.de) using the Biosemi Active2 system (www.biosemi.com). Electrodes were localized in three dimensions (Polhemus Patriot digitizer, www.polhemus.com), average referenced and band-pass filtered offline between 0.1 (12 dB/octave zero phase) and 200 Hz (24 dB/octave zero phase). EEG data were screened for electrodes with intermittent connectivity or epochs of extreme amplitude, and epochs were excluded from analyses if an eyeblink occurred coincident with stimulus presentation (between 150 msec prestimulus to 400 msec poststimulus onset), leaving an average of 62.4 hit trials (SD = 15.7, range = 22–92) and 742.4 correct rejection trials (SD = 80.8, range = 460–859) per participant. Second-order blind source identification (Belouchrani, Abed-Meriam, Cardoso, & Moulines, 1997) was then used to remove the remaining nonneural signal contaminants (i.e., 60 Hz contamination, ocular and muscle artifacts) from the trial-segmented EEG used to derive ERPs with a novel semiautomatic artifact removal tool (see Saggar et al., 2012). Finally, the 88-channel EEG data were reconstructed absent putative sources of noise and transformed using spherical spline interpolation (Perrin, Pernier, Bertrand, & Echallier, 1989) into a standard 81-channel montage (international 10–10 system) to ensure consistent channel locations across participants using BESA software (BESA 5.3; www.besa.de).

ERP Analysis

Modulations in the strength of the event-related electric field were quantified as GFP using the CARTOOL software package (Brunet, Murray, & Michel, 2011). GFP is a reference-independent measure of voltage potential (μV) that quantifies the strength of synchronized brain response, equivalent to the standard deviation of amplitude across the entire average reference electrode montage at a given sample of recording (Skrandies, 1990). GFP was calculated for hits and correct rejections, separately, from each participant's condition-averaged ERP (100 msec prestimulus to 900 msec poststimulus onset) baseline corrected to a 100-msec prestimulus period. Fifty-one participants (26 training) were included in Retreat 1 analyses, and 27 participants were included in Retreat 2 analyses, following listwise deletion due to missing data or poor recording quality at one or more assessments.

Modulations in electric field topography were examined using nonparametric randomization statistics implemented in Ragu (Koenig, Kottlow, Stein, & Melie-García, 2011). Ragu employs a procedure known as topographic ANOVA (TANOVA; Murray et al., 2008) by permuting an empirical distribution of global dissimilarity values. Global dissimilarity is quantified as the square root of the mean of the squared differences between each GFP normalized electrode and reflects a measure of the spatial configuration differences between two electric fields, independent of their strength. TANOVA compares the global dissimilarity of the original condition-averaged ERPs to the distribution of 5000 random permuted dissimilarities. Significant time windows can then be identified as values that are extreme compared with the permuted distribution (i.e., p values defined as 1 minus the proportion of permutations ≤ the actual global dissimilarity), which, because of physical laws (Vaughan, 1982), must reflect differences in the underlying configuration of intracranial generators (Murray et al., 2008).

Timewise ANOVA (least squares) and TANOVA (nonparametric permutation-based) were conducted on each sample of the condition-averaged event-related electric field to identify sustained time windows showing statistically significant differences in field strength or field topography, respectively. Timewise analysis provides an unbiased approach for assessing global modulations in the scalp electric field (Murray et al., 2008), absent need for experimenter-defined time windows or a priori electrode selection (Luck & Gaspelin, 2017). To account for temporal autocorrelation (cf. Guthrie & Buchwald, 1991), statistically significant time windows were defined as any range for which a p value of ≤.05 was sustained for at least 20 contiguous milliseconds. In addition, we confirmed that all reported timewise effects were sustained longer than the minimum critical threshold for contiguous significant samples described under the null hypothesis based on nonparametric permutations (Koenig et al., 2011). Retreat 1 analyses tested main effects of group (training and control), assessment (pre-, mid-, and postasssesment), and their interaction; Retreat 2 analyses tested effects of assessment only.

Post hoc analyses described the magnitude and direction of all sustained timewise effects. First, mixed-design ANOVA with Greenhouse–Geisser correction was applied to the mean GFP from each statistically significant time window. Pairwise differences were evaluated by comparison of marginal means. In addition, mixed-design ANOVA were used to characterize the distribution of significant effects in the electrode montage for each electrode from each time window in which we observed significant modulations in GFP or topography. Resultant p values from tests of the mean voltage across all 81 electrodes were subjected to false discovery rate (FDR) control of Type I error at the nominal level of α = .05 (Benjamini & Hochberg, 1995).

Vigilance Analysis

Within-task changes in GFP amplitude were examined by constructing condition averages of correct rejections for each of eight contiguous 4-min trial blocks (108 possible nontarget trials per block). Mixed-design ANOVA—with Block as a within-subject factor—were conducted on each GFP sample to identify statistically significant epochs lasting more than 20 msec. A total of 50 participants (25 training) were available for Retreat 1 analyses, and 25 participants were available for Retreat 2 analyses, following listwise deletion for missing data or poor recording quality at one or more blocks. Hits occurred too infrequently to conduct reliable block analyses.

Multilevel linear growth curves (Ferrer & McArdle, 2010) were used to model changes in mean GFP across blocks for each statistically significant epoch. Trajectories are described in terms of an intercept (i.e., starting point) and linear slope (i.e., rate of change). Fixed effects of Block and Assessment reflect the linear rate of change across each 4-min task segment and testing point, respectively, with random effects representing between-person variability in these parameters. All models were estimated using full maximum likelihood, implemented with SAS PROC MIXED Version 9.4. Participants excluded from timewise ANOVA were included in growth curve analyses, which can accommodate data missing at random.

RESULTS

We first examined grand-averaged electrode voltage waveforms and accompanying scalp topographies to characterize visual ERP components and their relation to correct discrimination between target and nontarget stimuli. ERPs from 81 electrode locations are depicted in Figure 2 for hits and correct rejections averaged across all participants (n = 51) and assessments in Retreat 1. These electrode voltage waveforms depict a clear succession of visual ERPs for both hits and correct rejections, beginning with a P1-like positivity over occipital electrode sites (peak positive voltage at PO3) occurring roughly 95–145 msec after stimulus onset, followed by an N1-like negativity over occipital electrode sites (peak negative voltage at PO8) occurring roughly 150–250 msec. A large P3-like positivity over central-parietal electrode sites (peak positive voltage at Pz) occurred later around 300–600 msec.

Figure 2. 

ERP waveforms from 81 electrode locations are depicted for hits and correct rejections (CR) averaged across all participants (n = 51) and assessments in Retreat 1. Condition-averaged activity is presented for the P1, N1, and P3 latency ranges. Black dots indicate electrodes demonstrating statistically significant paired voltage differences between grand-averaged condition ERPs following FDR correction. White dots indicate electrodes with peak voltage activity during each respective time range. Scalp voltage topographies are 2-D isometric projections with nasion upwards. A digital low-pass filter (30.0 Hz, 12 dB/octave) was applied before plotting waveforms in all figures.

Figure 2. 

ERP waveforms from 81 electrode locations are depicted for hits and correct rejections (CR) averaged across all participants (n = 51) and assessments in Retreat 1. Condition-averaged activity is presented for the P1, N1, and P3 latency ranges. Black dots indicate electrodes demonstrating statistically significant paired voltage differences between grand-averaged condition ERPs following FDR correction. White dots indicate electrodes with peak voltage activity during each respective time range. Scalp voltage topographies are 2-D isometric projections with nasion upwards. A digital low-pass filter (30.0 Hz, 12 dB/octave) was applied before plotting waveforms in all figures.

Paired t tests of the average amplitude at the P1 latency (95–145 msec) revealed 0 electrodes that significantly (FDR threshold of p < .034) differentiated hits from correct rejections, whereas 26 electrodes (FDR threshold < .016) were significantly different between conditions at the N1 latency (150–250 msec) with the largest negative mean difference (Mdiff = −0.475 μV) at electrode PO8. Also, 51 electrodes (FDR threshold < .027) differentiated conditions at the P3 latency (300–600 msec) with the largest positive mean difference (Mdiff = 1.023 μV) at electrode Pz. Timewise topographic comparison with TANOVA confirmed significant topographic differences between hits and correct rejections beginning 215 msec poststimulus to the end (900 msec) of the recorded epoch. These comparisons support the supposition that the N1 latency is a critical processing stage for stimulus discrimination and suggest that correctly detected targets are distinguished neurally from nontargets beginning at the latency of the N1, and are further differentiated in the stimulus processing stream by the activity of distinct neural generators post-N1.

Retreat 1

GFP and Topographic Modulations for Hits

Timewise analysis of GFP waveforms revealed temporally sustained and statistically significant main effects of assessment in the 410–560 msec range, and timewise analysis of the electric field topography with TANOVA revealed significant main effects of Group in the 215–285 msec range and significant effects of Assessment in the 400–420 msec range. There were no other significant effects. Post hoc analyses of global modulations in the strength (GFP) and topography of the electric field are summarized below for each significant time range, alongside comparisons of the electric field at the level of individual electrode voltage potentials. Group-averaged GFP waveforms, scalp voltage topographies, and significant electrode sites for all significant windows are depicted in Figure 3. These findings suggest a reduction from pre- to midassessment in synchronized activity and changes in voltage topography at P3 latencies (400–560 msec) regardless of group, coincident with the procedural increase in discrimination difficulty at midassessment. There were also group differences in the post-N1 (215–285 msec) topography, suggesting groups were distinguished based on activity of unique neural generators at this latency.

Figure 3. 

Mean GFP for hits for training (n = 26) and control (n = 25) participants at pre- (Pre), mid- (Mid), and postassessment (Post) in Retreat 1. Significant periods of sustained effects of Assessment in GFP are indicated by black bars along the x-axis. The bar graph (inset) depicts means and standard errors for mean GFP in the window of significant Assessment effects. Condition average voltage topographies are shown for epochs demonstrating significant Topographic (215–285 and 400–420 msec) or GFP (410–560 msec) effects between groups or assessments, respectively. Black dots indicate electrodes with significant effects of Assessment following FDR correction. Waveforms are depicted for exemplar electrodes indicated with a white dot. **p < .01.

Figure 3. 

Mean GFP for hits for training (n = 26) and control (n = 25) participants at pre- (Pre), mid- (Mid), and postassessment (Post) in Retreat 1. Significant periods of sustained effects of Assessment in GFP are indicated by black bars along the x-axis. The bar graph (inset) depicts means and standard errors for mean GFP in the window of significant Assessment effects. Condition average voltage topographies are shown for epochs demonstrating significant Topographic (215–285 and 400–420 msec) or GFP (410–560 msec) effects between groups or assessments, respectively. Black dots indicate electrodes with significant effects of Assessment following FDR correction. Waveforms are depicted for exemplar electrodes indicated with a white dot. **p < .01.

GFP 410–560 msec.

Post hoc analysis of the mean GFP in this late latency window revealed a significant effect of Assessment, F(2, 98) =6.53, p = .002, ηp2 = .118, but no significant effect of Group, F(1, 49) = .022, p = .882, ηp2 = .001, and no interaction between Group and Assessment, F(2, 98) = 1.07, p = .348, ηp2 = .021. GFP amplitude decreased significantly (p = .004) from preassessment (M = 3.91 μV, SE = 0.29) to midassessment (M = 3.34 μV, SE = 0.27) and decreased (p = .002) from pre- to postassessment (M = 3.29 μV, SE = 0.25) in all participants, but not from mid- to postassessment (p = .798). These modulations were reflected in 28 electrodes with significant ANOVA effects of Assessment (FDR threshold of p < .017, ηp2 range = .079–.273) within this window over central-parietal and occipital-temporal scalp locations with the largest effect at electrode PO10.

Topography 215–285 msec.

Follow-up analysis of the average topography in this window revealed a significant effect of Group (p = .001), but no significant effect of Assessment (p = .824) or interaction between Assessment and Group (p = .144). Zero electrodes had significant ANOVA effects of Group (FDR threshold < .007) with the strongest effect (p = .007, ηp2 = .138) at electrode AFz.

Topography 400–420 msec.

There was no significant effect of Group (p = .270), a significant effect of Assessment (p = .025), but no interaction between Assessment and Group (p = .589). Topography significantly differed (p = .048) between pre- and midassessment, differed (p = .034) between pre- and postassessment, but did not differ (p = .226) between mid- and postassessment. Twenty-one electrodes had significant effects of Assessment (FDR threshold < .011, ηp2 range = .092–.228) over central-parietal and occipital-temporal scalp locations with the largest effect at PO10.

GFP and Topographic Modulations for Correct Rejections

Timewise analysis of GFP waveforms confirmed temporally sustained and statistically significant ANOVA effects of Assessment in the 425–500 msec window, and timewise analysis of the electric field topography with TANOVA revealed significant main effects of Assessment in the 240–275 and 330–460 msec ranges. Post hoc analyses are summarized below for each significant time range, alongside comparisons at the level of individual electrode voltage potentials. Group-averaged GFP waveforms, scalp voltage topographies, and significant electrode sites for all significant windows are depicted in Figure 4. Consistent with the pattern for hits, GFP amplitude for correct rejections also decreased and accompanied topographic changes from pre- to midassessment at P3 latencies across groups, coincident with the increase in discrimination difficulty.

Figure 4. 

Mean GFP for correct rejections (CR) for training (n = 26) and control (n = 25) participants at pre- (Pre), mid- (Mid), and postassessment (Post) in Retreat 1. Significant periods of sustained effects of Assessment in GFP are indicated by black bars along the x-axis. The bar graph (inset) depicts means and standard errors for mean GFP in the window of significant Assessment effects. Condition average voltage topographies are shown for epochs demonstrating significant Topographic (240–275 and 330–460 msec) or GFP (425–500 msec) effects between Assessments. Black dots indicate electrodes with significant effects of Assessment following FDR correction. Waveforms are depicted for exemplar electrodes indicated with a white dot. **p < .01.

Figure 4. 

Mean GFP for correct rejections (CR) for training (n = 26) and control (n = 25) participants at pre- (Pre), mid- (Mid), and postassessment (Post) in Retreat 1. Significant periods of sustained effects of Assessment in GFP are indicated by black bars along the x-axis. The bar graph (inset) depicts means and standard errors for mean GFP in the window of significant Assessment effects. Condition average voltage topographies are shown for epochs demonstrating significant Topographic (240–275 and 330–460 msec) or GFP (425–500 msec) effects between Assessments. Black dots indicate electrodes with significant effects of Assessment following FDR correction. Waveforms are depicted for exemplar electrodes indicated with a white dot. **p < .01.

GFP 425–500 msec.

Post hoc analyses on the mean GFP in this window revealed a significant effect of Assessment, F(2, 98) = 5.690, p = .006, ηp2 = .104, but no significant effect of Group, F(1, 49) = 1.828, p = .183, ηp2 = .036, and no interaction between Group and Assessment, F(2, 98) = .907, p = .398, ηp2 = .018. GFP amplitude decreased significantly (p = .005) from pre- (M = 2.61 μV, SE = 0.20) to midassessment (M = 2.29 μV, SE = 0.17) and decreased (p = .007) from pre- to postassessment (M = 2.20 μV, SE = 0.17) in both groups but did not change from mid- to postassessment (p = .482). These modulations reflected 35 electrodes with significant effects of Assessment (FDR threshold of p < .021, ηp2 range = .079–.255) within this window over central and occipital-temporal locations with the largest effect at Cz.

Topography 240–275 msec.

Analysis of the average topographies within this window revealed a significant effect of Group (p = .005), a significant effect of Assessment (p = .029), but no interaction between Assessment and Group (p = .679). There were, however, no significant differences between any assessment (all ps > .058), and zero electrodes with significant ANOVA effects of Assessment (FDR threshold < .011) with the strongest effect (p = .011, ηp2 = .087) at electrode P6.

Topography 330–460 msec.

There was no significant effect of Group (p = .288), a significant effect of Assessment (p < .001), and no significant interaction between Assessment and Group (p = .777). Topography significantly differed (p = .024) between pre- and midassessment, differed (p < .001) between pre- and postassessment, but did not differ (p = .052) between mid- and postassessment. Thirty-six electrodes had significant ANOVA effects of Assessment (FDR threshold < .018, ηp2 range = .078–.295) over central and occipital-temporal scalp locations with the largest effect at electrode Cz.

GFP Correlates of Visual Perceptual Threshold

We previously reported that training participants showed greater visual discrimination capacity (perceptual threshold) following training than did wait-list controls (MacLean et al., 2010). A significant Assessment × Group interaction was also observed in the present EEG subsample, F(2, 98) = 3.540, p = .036, ηp2 = .067. Training group participants had significantly greater (p = .012) discrimination threshold than controls at postassessment, but not at pre- or midassessment (ps > .237). Perceptual discrimination threshold thus increased more for the training participants than controls (see descriptive statistics in Table 1).

We next examined whether individual differences in target stimulus line length—which varied according to an individual's discrimination threshold—predicted GFP amplitude of visual ERPs during the CPT. Timewise ANCOVA analyses (with effects of Target line length, Group, and their interaction as covariates) revealed a large number of temporally sustained (>20 msec) correlations between Target line length and GFP amplitude at mid- and postassessment for both hits and correct rejections. For each significant epoch, we first evaluated the effect of Target line length on mean GFP, then included Group, and finally the interaction between Target length and Group as sequential regression predictors to examine the direction and magnitude of effects. Table 2 presents the unique variance explained by each regression step and zero-order correlations for each group in each range of significant effects. Scatterplots for each epoch are depicted in Figure 5.

Table 2. 
Discrimination Threshold Predicts GFP Amplitude in Retreat 1
Epoch (msec)Step 1: ThresholdStep 2: GroupStep 3: InteractionZero-order Correlation (r)
R2 (1, 49)ΔR2 (1, 48)ΔR2 (1, 47)TrainingControl
Midassessment 
 140–170, Hit .129** .098* .005 .45* .41* 
 255–290, Hit .014 .005 .130** .51** −.10 
 175–230, CR .121* .010 .020 .39* .32 
 250–285, CR .032 .001 .134** .52** −.04 
 320–360, CR .076* .004 .098* .62** .12 
 345–375, CR .091* .001 .081* .56** .17 
  
Postassessment 
 175–195, Hit .108* .002 .001 .27 .42* 
 155–195, CR .121* .004 .007 .35 .38 
Epoch (msec)Step 1: ThresholdStep 2: GroupStep 3: InteractionZero-order Correlation (r)
R2 (1, 49)ΔR2 (1, 48)ΔR2 (1, 47)TrainingControl
Midassessment 
 140–170, Hit .129** .098* .005 .45* .41* 
 255–290, Hit .014 .005 .130** .51** −.10 
 175–230, CR .121* .010 .020 .39* .32 
 250–285, CR .032 .001 .134** .52** −.04 
 320–360, CR .076* .004 .098* .62** .12 
 345–375, CR .091* .001 .081* .56** .17 
  
Postassessment 
 175–195, Hit .108* .002 .001 .27 .42* 
 155–195, CR .121* .004 .007 .35 .38 

R2 values (degrees of freedom) and correlation coefficients are reported for significant epochs identified by timewise regression, separately for midassessment (Mid) and postassessment (Post). R2 is the variance explained by target length (threshold) as a single predictor; ΔR2 is the unique variance explained by the additional predictors of Group (training vs. control) and the Group × Target Length interaction, entered as sequential regression steps. CR = correct rejection.

*

p < .05.

**

p < .01.

Figure 5. 

Scatterplots of mean GFP by target length (PEST discrimination threshold in degrees of visual angle) for hits and correct rejections (CR) for significant epochs in Retreat 1.

Figure 5. 

Scatterplots of mean GFP by target length (PEST discrimination threshold in degrees of visual angle) for hits and correct rejections (CR) for significant epochs in Retreat 1.

Target line length positively predicted GFP amplitude in the period of 140–230 msec after stimulus onset for both hits and correct rejections (R2 range = .108–.129). Thus, individual differences in target length predicted GFP amplitude to nontargets, which were held at constant length for all participants across assessments—a pattern suggestive of top–down influences of target discrimination difficulty on perceptual processing of nontargets. For the 250–290 msec period, discrimination threshold was positively related to both target and nontarget amplitude in training participants (ΔR2 range = .130–.134, for Target Length × Group interaction), but not in control participants (all ps > .46). Overall, these analyses suggest that response strength at N1 latencies (∼150–220 msec) were sensitive to individual differences in target stimulus length across groups. Response strength was correlated with target length significantly more for training participants than controls at subsequent epochs (∼250–375 msec) at latencies in which we observed topographic group differences in hits.

Vigilance Analysis

Visual inspection of GFP waveforms for correct rejections suggested a monotonic decline in amplitude across contiguous task blocks. Timewise ANOVA confirmed a temporally sustained main effect of Block in the range of 150–245 msec. No interactions of Block with Group or Assessment were observed. Figure 6 depicts this within-task decline in GFP amplitude, averaged across assessments, separately for each group.

Figure 6. 

Mean GFP for correct rejections, averaged across Retreat 1 assessments for training (n = 26) and control (n = 25) participants from 0 to 300 msec poststimulus onset. Each waveform is the average of all correct rejections within each contiguous 4-min task block. Significant periods of sustained effects of Block are indicated by black bars along the x-axis. The line graph (inset) depicts mean GFP in the window of significant Block effects (150–245 msec) across blocks for training (red) and control groups (black). Grand-averaged scalp voltage topography is depicted for the significant epoch identified from timewise analysis of the GFP. Black dots indicate electrodes with significant linear effects of Block (ps < .034) following FDR correction.

Figure 6. 

Mean GFP for correct rejections, averaged across Retreat 1 assessments for training (n = 26) and control (n = 25) participants from 0 to 300 msec poststimulus onset. Each waveform is the average of all correct rejections within each contiguous 4-min task block. Significant periods of sustained effects of Block are indicated by black bars along the x-axis. The line graph (inset) depicts mean GFP in the window of significant Block effects (150–245 msec) across blocks for training (red) and control groups (black). Grand-averaged scalp voltage topography is depicted for the significant epoch identified from timewise analysis of the GFP. Black dots indicate electrodes with significant linear effects of Block (ps < .034) following FDR correction.

The magnitude of vigilance effects in the 150–245 msec window was examined via growth curve analysis. Multilevel models revealed a significant effect of Block, F(1, 1166) = 27.68, p < .001, and a significant interaction between Group and Assessment, F(1, 1166) = 6.24, p = .013, but no other main effects or interactions. Parameter estimates from a simplified model including only main effects and the Assessment × Group interaction are given in Table 3. The significant linear effect of block indicates that GFP amplitude declined by −0.045 μV (p < .001) across each 4-min task block in this epoch. In addition, the significant interaction between Group and Assessment suggests that groups changed differently in mean amplitude across assessments. Control participants increased, on average, by 0.063 μV (p = .041) per assessment, whereas training participants decreased marginally across assessments by −0.058 μV (p = .055); nevertheless, there were no significant group differences at any assessment (all ps > .150). Follow-up growth curve analyses of the electrode montage revealed 55 electrodes with significant linear effects of block (FDR threshold < .034) in the epoch of significant GFP effects (see Figure 6).

Table 3. 
Parameter Estimates from Model of Within-task Change in GFP Amplitude for Retreat 1
Model EffectsParameter Estimate (SE)
150–245 msec
Fixed effects 
 Intercept 1.755 (0.127)*** 
 Group 0.235 (0.160) 
 Linear block −0.045 (0.006)*** 
 Linear assessment 0.063 (0.031) 
 Group × Assessment −0.122 (0.043)* 
  
Random effects 
 Intercept variance, σ02 0.444 (0.094)*** 
 Block variance, σb2 0.001 (0.001) 
 Assessment variance, σa2 0.014 (0.005)** 
 Intercept-block, σ0,b −0.011 (0.004)* 
 Intercept-assessment, σ0,a −0.012 (0.015) 
 Block-assessment, σb,a −0.001 (0.001) 
 Residual variance, σe2 0.158 (0.007)*** 
  
−2 Log-likelihood 1464.5 
Model EffectsParameter Estimate (SE)
150–245 msec
Fixed effects 
 Intercept 1.755 (0.127)*** 
 Group 0.235 (0.160) 
 Linear block −0.045 (0.006)*** 
 Linear assessment 0.063 (0.031) 
 Group × Assessment −0.122 (0.043)* 
  
Random effects 
 Intercept variance, σ02 0.444 (0.094)*** 
 Block variance, σb2 0.001 (0.001) 
 Assessment variance, σa2 0.014 (0.005)** 
 Intercept-block, σ0,b −0.011 (0.004)* 
 Intercept-assessment, σ0,a −0.012 (0.015) 
 Block-assessment, σb,a −0.001 (0.001) 
 Residual variance, σe2 0.158 (0.007)*** 
  
−2 Log-likelihood 1464.5 

Maximum likelihood estimates are reported for correct rejection GFP amplitude in Retreat 1. Standard errors are reported in parentheses.

*

p < .05.

**

p < .01.

***

p < .001.

Retreat 2

During Retreat 1, CPT target length was individually adjusted at each assessment to match changes in participants' perceptual thresholds. In Retreat 2 (n = 27), we examined changes across assessments in the event-related neuroelectric field when target stimulus length was instead held constant in the CPT and participants' perceptual threshold could vary relative to target length. To provide confirmatory within-subject support, we also conducted directed comparisons of GFP and topography between pre- and postassessment in Retreat 1 and Retreat 2 in the subset of participants with complete data (n = 22) across all six study assessments.

GFP and Topographic Modulations for Hits

Timewise analyses of GFP waveforms revealed temporally sustained effects of Assessment in the range of 455–520 msec. Analysis of the electric field topography also revealed significant main effects of Assessment in the 145–175 msec range. Post hoc analyses are summarized below. Group-averaged GFP waveforms, scalp voltage topographies, and significant electrode sites for all significant windows are depicted in Figure 7. As can be seen, effects in the 455–520 msec range reflect increases in amplitude of the electric field, whereas topographic effects at the N1 onset latency range (145–175 msec) suggest earlier onset of the voltage map configuration over assessments.

Figure 7. 

Mean GFP for hits for wait-list training group participants (n = 27) at pre- (Pre), mid- (Mid), and postassessment (Post) in Retreat 2, with GFP for controls at Retreat 1 postassessment (n = 22) shown for reference. Significant periods of sustained effects of Assessment in Retreat 2 are indicated by black bars along the x-axis. The bar graph (inset) depicts means and standard errors for mean GFP in the window of significant Assessment effects. Condition average voltage topographies are shown for epochs demonstrating significant Topographic (145–175 msec) or GFP (455–520 msec) effects between assessments. Black dots indicate electrodes with significant effects of Assessment following FDR correction. Waveforms are depicted for exemplar electrodes indicated with a white dot. *p < .05.

Figure 7. 

Mean GFP for hits for wait-list training group participants (n = 27) at pre- (Pre), mid- (Mid), and postassessment (Post) in Retreat 2, with GFP for controls at Retreat 1 postassessment (n = 22) shown for reference. Significant periods of sustained effects of Assessment in Retreat 2 are indicated by black bars along the x-axis. The bar graph (inset) depicts means and standard errors for mean GFP in the window of significant Assessment effects. Condition average voltage topographies are shown for epochs demonstrating significant Topographic (145–175 msec) or GFP (455–520 msec) effects between assessments. Black dots indicate electrodes with significant effects of Assessment following FDR correction. Waveforms are depicted for exemplar electrodes indicated with a white dot. *p < .05.

GFP 455–520 msec.

ANOVA analyses confirmed a significant effect of Assessment for mean GFP amplitude in this epoch, F(2, 52) = 4.309, p = .023, ηp2 = .142. Amplitude significantly (p = .019) increased from pre- (M = 3.41 μV, SE = 0.21) to postassessment (M = 4.08 μV, SE = 0.35) and increased (p = .049) from mid- (M = 3.64 μV, SE = 0.32) to postassessment. Preassessment did not differ (p = .279) from midassessment. These effects were reflected in 18 electrodes with significant ANOVA effects of Assessment (FDR threshold of p < .011, ηp2 range = .166–.269) at frontocentral and occipital-temporal scalp locations with the largest effect at electrode Oz. Finally, we examined change in GFP amplitude at this epoch from pre- to postassessment in Retreat 1 and Retreat 2 in the subset of participants with complete data (n = 22). Amplitude significantly (p = .019) decreased from pre- to postassessment (Mdiff = −1.14 μV, SE = 0.28) in Retreat 1, but marginally increased (p = .052) from pre- to postassessment (Mdiff = 0.70 μV, SE = 0.34) in Retreat 2.

Topography 145–175 msec.

Analysis of the average topography in this range revealed a significant effect of Assessment (p = .001). Topography significantly differed (p = .001) between pre- and midassessment, differed (p = .002) between pre- and postassessment, but did not differ (p = .739) between mid- and postassessment. Eight electrodes had significant effects of Assessment (FDR threshold of p < .004, ηp2 range = .198–.325) over bilateral-frontal and occipital-parietal scalp locations with the largest effect at P4. Comparing across retreats (n = 22), topography did not change (p = .306) from pre- to postassessment in Retreat 1 but changed (p = .003) from pre- to postassessment in Retreat 2.

GFP and Topographic Modulations for Correct Rejections

Timewise analysis of GFP waveforms for correct rejections revealed a significant effect of Assessment in the 155–205 msec range. Analysis of the electric field topography also revealed significant main effects of Assessment in the 90–145, 150–175, and 270–330 msec ranges. Post hoc analyses are summarized below. Group-averaged GFP waveforms, scalp voltage topographies, and significant electrode sites for all significant windows are depicted in Figure 8. Effects in the N1 latency range (150–175 and 155–205 msec ranges) reflect increases in amplitude and faster onset of the scalp voltage map representing the N1 across assessments, and topographic modulations in the 90–145 msec range reflect faster onset of the P1-like voltage map across assessments.

Figure 8. 

Mean GFP for correct rejections (CR) for wait-list training group participants (n = 27) at pre- (Pre), mid- (Mid), and postassessment (Post) in Retreat 2, with GFP for controls at Retreat 1 postassessment (n = 22) shown for reference. Significant periods of sustained effects of Assessment in Retreat 2 are indicated by black bars along the x-axis. The bar graph (inset) depicts means and standard errors for mean GFP in the window of significant Assessment effects. Condition average voltage topographies are shown for epochs demonstrating significant Topographic (90–145, 150–175, and 270–330 msec) or GFP (155–205 msec) effects between assessments. Black dots indicate electrodes with significant effects of Assessment following FDR correction. Waveforms are depicted for exemplar electrodes indicated with a white dot. **p < .01.

Figure 8. 

Mean GFP for correct rejections (CR) for wait-list training group participants (n = 27) at pre- (Pre), mid- (Mid), and postassessment (Post) in Retreat 2, with GFP for controls at Retreat 1 postassessment (n = 22) shown for reference. Significant periods of sustained effects of Assessment in Retreat 2 are indicated by black bars along the x-axis. The bar graph (inset) depicts means and standard errors for mean GFP in the window of significant Assessment effects. Condition average voltage topographies are shown for epochs demonstrating significant Topographic (90–145, 150–175, and 270–330 msec) or GFP (155–205 msec) effects between assessments. Black dots indicate electrodes with significant effects of Assessment following FDR correction. Waveforms are depicted for exemplar electrodes indicated with a white dot. **p < .01.

GFP 155–205 msec.

ANOVA analyses confirmed a significant effect of Assessment for mean GFP amplitude in this window, F(2, 52) = 8.04, p = .003, ηp2 = .236. Amplitude significantly increased (p = .003) from preassessment (M = 1.26 μV, SE = 0.10) to midassessment (M = 1.42 μV, SE = 0.11) and increased (p = .003) from pre- to postassessment (M = 1.52 μV, SE = 0.14). Midassessment did not differ (p = .125) from postassessment. These effects were reflected in 16 electrodes with significant effects of Assessment (FDR threshold of p < .010, ηp2 range = .176–.325) over frontal and occipital-parietal locations with the largest effect at electrode PO7. Comparing across retreats (n = 22), amplitude did not change (p = .378) from pre- to postassessment (Mdiff = −0.07 μV, SE = 0.08) in Retreat 1 but significantly increased (p = .024) from pre- to postassessment (Mdiff = 0.19 μV, SE = 0.08) in Retreat 2.

Topography 90–145 msec.

There was a significant effect of Assessment (p < .001) for the average topography within this window. Topography significantly differed (p < .001) between pre- and midassessment, differed (p = .019) between pre- and postassessment, but did not differ (p = .057) between mid- and postassessment. Twenty-six electrodes had significant effects of Assessment (FDR threshold of p < .015, ηp2 range = .152–.325) over frontal and occipital-parietal scalp locations with the largest effect at electrode F4. Comparing across retreats (n = 22), topography did not change (p = .524) from pre- to postassessment in Retreat 1 but significantly changed (p = .011) from pre- to postassessment in Retreat 2.

Topography 150–175 msec.

There was a significant effect of Assessment (p = .004) for the average topography within this window. Topography significantly differed (p = .002) between pre- and midassessment, differed (p = .006) between pre- and postassessment, but did not differ (p = .901) between mid- and postassessment. Eleven electrodes had significant effects of Assessment (FDR threshold of p < .006, ηp2 range = .187–.292) over frontal and occipital-parietal scalp locations with the largest effect at F7. Comparing across retreats (n = 22), topography did not change (p = .389) from pre- to postassessment in Retreat 1 but significantly changed (p = .028) from pre- to postassessment in Retreat 2.

Topography 270–330 msec.

There was a significant effect of Assessment (p = .003) for the average topography within this window. Topography significantly differed (p = .011) between pre- and midassessment, differed (p = .010) between pre- and postassessment, but did not differ (p = .887) between mid- and postassessment. Six electrodes had significant effects of Assessment (FDR threshold of p < .004, ηp2 range = .219–.313) over frontocentral scalp locations with the largest effect at Cz. Comparing across retreats (n = 22), topography did not change (p = .272) from pre- to postassessment in Retreat 1 but significantly changed (p = .036) from pre- to postassessment in Retreat 2.

GFP Correlates of Discrimination Capacity

Consistent with MacLean et al. (2010), the present subsample of EEG participants demonstrated an increased discrimination threshold across the Retreat 2 training period, F(2, 52) = 7.496, p = .002, ηp2 = .224 (see Table 1 for descriptive statistics). Individuals' discrimination threshold significantly increased (p = .009) from pre- to midassessment and increased (p = .002) from pre- to postassessment, but not from mid- to postassessment (p = .675).3 In exploratory analyses, we examined whether electric field strength was influenced by the difference between each participant's measured discrimination threshold at the mid- or postassessment and their task set target length (determined by their achieved discrimination threshold from preassessment). At mid- and postassessment, timewise regression analyses revealed a temporally sustained effect of threshold difference (measured threshold –target line length) on GFP amplitude for hits and correct rejections.4

Linear regressions for each significant timewise epoch are reported in Table 4, with scatterplots for each epoch depicted in Figure 9. The discrepancy between participants' measured perceptual threshold and target line length explained a large amount of variance in GFP in the 350–860 msec window for midassessment hits, R2 = .382, F(1, 25) = 15.48, p < .001, and in the 395–640 msec epoch for postassessment hits, R2 = .303, F(1, 25) = 10.89, p = .003, such that GFP amplitude was predicted to increase by 4.2 μV at midassessment (p = .001) and 3.3 μV at postassessment (p = .003) for each 1° of visual angle that perceptual threshold exceeded the target line length. A similar pattern was observed for correct rejections at postassessment, for which threshold difference was a significant predictor of mean GFP in the 460–485 msec epoch, R2 = .257, F(1, 25) = 8.64, p = .007.

Table 4. 
Discrimination Capacity (Δ Threshold) Predicts GFP Amplitude in Retreat 2
Epoch (msec)R2 (1, 25)Correlation (r)
Midassessment 
 350–860, Hit .382*** .62*** 
  
Postassessment 
 395–640, Hit .303** .55** 
 460–485, CR .257** .51** 
Epoch (msec)R2 (1, 25)Correlation (r)
Midassessment 
 350–860, Hit .382*** .62*** 
  
Postassessment 
 395–640, Hit .303** .55** 
 460–485, CR .257** .51** 

R2 values (degrees of freedom) and correlation coefficients are reported for significant epochs identified by timewise regression, separately for midassessment (Mid) and postassessment (Post). R2 is the variance explained by the difference (Δ) between a participants' discrimination threshold at mid- or postassessment and their PEST target length at preassessment. CR = correct rejection.

*

p < .05.

**

p < .01.

***

p < .001.

Figure 9. 

Scatterplots of mean GFP by Δ discrimination for hits and correct rejections (CR) for significant epochs in Retreat 2, where Δ discrimination reflects the observed difference between a participant's PEST target length at mid- (Mid) or postassessment (Post) and their PEST target length at preassessment (in degrees of visual angle).

Figure 9. 

Scatterplots of mean GFP by Δ discrimination for hits and correct rejections (CR) for significant epochs in Retreat 2, where Δ discrimination reflects the observed difference between a participant's PEST target length at mid- (Mid) or postassessment (Post) and their PEST target length at preassessment (in degrees of visual angle).

Overall, individuals had greater GFP amplitude in later epochs when their perceptual threshold exceeded the difficulty of the target discrimination, suggesting that relative discrimination capacity was a strong correlate of later stimulus processing in the CPT. To rule out the influence of RT-related motor potentials on GFP amplitude at later epochs, we included participants' median target RT as an additional covariate in timewise regressions for hits. Median RT did not predict GFP amplitude for any sustained epoch (no ps < .05 for > 20 msec), or significantly alter the relation between threshold difference and GFP amplitude when included as a covariate.

Vigilance Analysis

As in Retreat 1, GFP amplitude appeared to decline with time on task at latencies associated with perceptual processing. Temporally sustained main effects of Block were found in the 150–250 and 365–595 msec windows, and significant interactions were found between Block and Assessment in the 175–225 and 545–590 msec windows. Figure 10 depicts these within-task decline in GFP for each assessment, along with the grand-averaged scalp topography.

Figure 10. 

Mean GFP for correct rejections at pre- (Pre), mid- (Mid), and postassessments (Post) in Retreat 2. Each waveform is the average of all correct rejections within each contiguous 4-min task block. Significant periods of sustained effects of Block are indicated by black bars, and significant interactions between Block and Assessment are indicated with checked bars along the x-axis. The line graph (inset) depicts mean GFP in the window of significant interactions between Block and Assessment (175–225 msec) at pre- (dotted), mid- (solid black), and postassessment (thick black). Grand-averaged scalp voltage topography is depicted for significant epochs of the GFP. Black dots indicate electrodes with significant interaction effects of Assessment and Block following FDR correction in the 175–225 msec (ps < .006) and 545–590 msec (ps < .008) epochs.

Figure 10. 

Mean GFP for correct rejections at pre- (Pre), mid- (Mid), and postassessments (Post) in Retreat 2. Each waveform is the average of all correct rejections within each contiguous 4-min task block. Significant periods of sustained effects of Block are indicated by black bars, and significant interactions between Block and Assessment are indicated with checked bars along the x-axis. The line graph (inset) depicts mean GFP in the window of significant interactions between Block and Assessment (175–225 msec) at pre- (dotted), mid- (solid black), and postassessment (thick black). Grand-averaged scalp voltage topography is depicted for significant epochs of the GFP. Black dots indicate electrodes with significant interaction effects of Assessment and Block following FDR correction in the 175–225 msec (ps < .006) and 545–590 msec (ps < .008) epochs.

Table 5 presents parameter estimates from growth curve models of the mean GFP for each significant epoch. In all cases, multilevel models revealed significant effects of Block, Fs(1, 615) > 14.39, ps < .001, and significant interactions between Block and Assessment, Fs(1, 615) > 6.58, ps < .011. Mean amplitude declined linearly at preassessment for each 4-min block, with the largest (β = −0.085 μV) and smallest (β = −0.063 μV) decline observed for the 545–590 and 150–250 msec windows, respectively. Importantly, the rate of within-task amplitude reduction was positively moderated across assessments for all significant windows, as indicated by the positive parameters for Block × Assessment. Furthermore, at later epochs, significant main effects of Assessment were also observed, Fs(1, 615) > 5.21, ps < .023, indicating a decrease in overall level of GFP amplitude across assessments. Finally, follow-up analyses of the electrode montage revealed nine electrodes with significant interactions of Block and Assessment (FDR threshold < .006) in the 175–225 msec window, and 11 electrodes with significant interactions of Block and Assessment (FDR threshold < .008) in the 545–590 msec window. For early latencies, within-task reductions in GFP amplitude were thus reduced in magnitude across assessments.

Table 5. 
Parameter Estimates from Models of Within-task Change in GFP Amplitude for Retreat 2
Model EffectsParameter Estimate (SE)
150–250 msec175–225 msec365–595 msec545–590 msec
Fixed effects 
 Intercept 1.900 (0.117)*** 2.081 (0.140)*** 2.648 (0.177)*** 2.636 (0.185)*** 
 Linear block −0.063 (0.009)*** −0.078 (0.010)*** −0.078 (0.016)*** −0.085 (0.022)*** 
 Linear assessment 0.005 (0.043) 0.003 (0.049) −0.169 (0.074)* −0.317 (0.090)*** 
 Block × Assessment 0.015 (0.006)* 0.021 (0.007)** 0.035 (0.011)** 0.057 (0.013)*** 
  
Random effects 
 Intercept variance, σ02 0.344 (0.098)*** 0.490 (0.139)*** 0.751 (0.220)*** 0.783 (0.235)*** 
 Block variance, σb2 0.001 (0.000)* 0.001 (0.001)* 0.001 (0.001) 0.005 (0.002)* 
 Assessment variance, σa2 0.034 (0.011)*** 0.044 (0.014)*** 0.091 (0.030)** 0.133 (0.044)** 
 Intercept-block, σ0,b −0.010 (0.005)* −0.014 (0.007)* −0.019 (0.011) −0.023 (0.018) 
 Intercept-assessment, σ0,a 0.013 (0.023) 0.017 (0.031) 0.006 (0.057) −0.041 (0.073) 
 Block-assessment, σb,a −0.002 (0.001) −0.002 (0.002) 0.001 (0.004) 0.011 (0.008) 
 Residual variance, σe2 0.072 (0.004)*** 0.100 (0.006)*** 0.262 (0.016)*** 0.398 (0.024)*** 
  
−2 Log-likelihood 323.6 535.7 1137.0 1411.6 
Model EffectsParameter Estimate (SE)
150–250 msec175–225 msec365–595 msec545–590 msec
Fixed effects 
 Intercept 1.900 (0.117)*** 2.081 (0.140)*** 2.648 (0.177)*** 2.636 (0.185)*** 
 Linear block −0.063 (0.009)*** −0.078 (0.010)*** −0.078 (0.016)*** −0.085 (0.022)*** 
 Linear assessment 0.005 (0.043) 0.003 (0.049) −0.169 (0.074)* −0.317 (0.090)*** 
 Block × Assessment 0.015 (0.006)* 0.021 (0.007)** 0.035 (0.011)** 0.057 (0.013)*** 
  
Random effects 
 Intercept variance, σ02 0.344 (0.098)*** 0.490 (0.139)*** 0.751 (0.220)*** 0.783 (0.235)*** 
 Block variance, σb2 0.001 (0.000)* 0.001 (0.001)* 0.001 (0.001) 0.005 (0.002)* 
 Assessment variance, σa2 0.034 (0.011)*** 0.044 (0.014)*** 0.091 (0.030)** 0.133 (0.044)** 
 Intercept-block, σ0,b −0.010 (0.005)* −0.014 (0.007)* −0.019 (0.011) −0.023 (0.018) 
 Intercept-assessment, σ0,a 0.013 (0.023) 0.017 (0.031) 0.006 (0.057) −0.041 (0.073) 
 Block-assessment, σb,a −0.002 (0.001) −0.002 (0.002) 0.001 (0.004) 0.011 (0.008) 
 Residual variance, σe2 0.072 (0.004)*** 0.100 (0.006)*** 0.262 (0.016)*** 0.398 (0.024)*** 
  
−2 Log-likelihood 323.6 535.7 1137.0 1411.6 

Maximum likelihood estimates are reported for correct rejection GFP amplitude in Retreat 2. Standard errors are reported in parentheses.

*

p < .05.

**

p < .01.

***

p < .001.

DISCUSSION

The present findings offer converging evidence for the moderating effects of perceptual discrimination ability on the maintenance and malleability of attentional performance over time. First, it appears that the contextual demand imposed by the difficulty of visual discrimination, when considered relative to one's capacity, can exert a systematic influence on neural activity at early processing stages with downstream consequences for further stimulus processing and vigilance. Second, improvements in perceptual discrimination associated with meditation seem to influence changes in the strength and topography of event-related neuroelectric activity at both earlier perceptual latencies as well as later stimulus processing stages, but only when discrimination capacity is allowed to exceed the difficulty imposed by the visual target.

As previously reported (MacLean et al., 2010), participants' ability to discriminate between line stimuli in a perceptual threshold procedure improved following two independent 3-month retreat interventions. In Retreat 1, discrimination thresholds were used to adjust the length of the CPT target stimulus across individuals and assessments. In line with previous behavioral findings (MacLean et al., 2010), no training-related changes in event-related brain activity were observed when CPT target length was adjusted to match changes in participants' current discrimination threshold. In contrast, both groups showed a reduction in P3-related brain response strength (GFP) when visual discriminations were made more demanding relative to preassessment. That no training-specific changes were observed, despite improvements in discrimination capacity, supports the idea that discrimination difficulty is a critical source of task interference and may function as a processing bottleneck for attentional modulation through meditation training.

In Retreat 2, CPT target length was held constant across all assessments, irrespective of changes in discrimination capacity. In this case, participants' visual discrimination improved, and we observed clear modulations in the scalp electric field at latencies typical of the visual N1 and P3. At earlier sensory latencies, these modulations of the electric field appear to reflect faster onset of voltage maps typical of the N1 for both hits and correct rejections, as well as faster P1 onset and greater N1-related GFP amplitude at N1 latencies for correct rejections alone. At later latencies, increases in P3-related GFP amplitude for hits were observed, suggesting that improvements in perceptual threshold may facilitate attentional processing of visual targets with intensive training, perhaps by enhancing the fidelity of stimulus representations for attentional selection. This notion was directly supported by strong correlations (up to 38% of variance explained for hits, 26% for correct rejections) between GFP amplitude at P3 latencies and the degree to which participants' discrimination threshold improved relative to preassessment. In addition, topographic changes, manifesting as increased negative voltage potential at central electrodes, were observed in correct rejections about 300 msec after stimulus presentation. Meditation training may therefore recruit distinct neural generators at this latency, perhaps reflecting the activity of template matching mechanisms (Folstein & Van Petten, 2008).

In both retreats, the attempt to sustain attention over minutes of performance exerted a systematic influence on GFP amplitude at early sensory latencies. The attentional demands of vigilance thus appeared to precipitate a degradation of synchronized neural activity at perceptual processing stages. These effects, however, may instead reflect the consequences of neural adaptation rather than diminished vigilance because we did not isolate attentional effects by comparing between stimulus conditions. This is perhaps unlikely given that within-task change was specific to the N1 latency but not earlier latencies. Future studies should explore this possibility by comparing relative degradation across stimulus conditions (i.e., relative to hits or misses) using more trials. Concordant with global task changes, we observed an attenuation of within-task declines in GFP amplitude across Retreat 2 assessments only, when target difficulty was fixed. It is thus possible that task-averaged ERP effects at the N1 latency resulted from more stable and consistent top–down attentional engagement with stimuli over the course of the task. One possibility, not addressed here, is that oscillatory mechanisms driving phase entrainment of attentional networks to ongoing stimulus input could have increased the timing consistency of evoked sensory responses with intensive training (Lutz et al., 2009; see also Slagter et al., 2011). An attenuation of amplitude decrements was also observed at delayed poststimulus periods in Retreat 2, suggestive of later consolidation or stimulus categorization processes (Polich, 2007). Overall, there are striking parallels between the attenuation of vigilance effects reported here and the moderation of behavioral decrements reported in MacLean et al. (2010), and links between neural and behavioral measures of vigilance should continue to be explored in future studies.

One interpretation of our results is that increased frontal P3 amplitude (i.e., P3a; Polich, 2007) in Retreat 2 is suggestive of enhanced exogenous orienting to more perceptually salient targets following intensive training. Such an interpretation aligns with the practice of shamatha in which the practitioner rests their attention on their meditative object in an increasingly effortless manner, allowing the ongoing flow of perceptions to act as an anchor for attention over time (Wallace, 2006). Bottom–up changes in perceptual capacities could have led to greater stimulus-driven orienting as practitioners' attention is likewise naturally anchored to a stream of more perceptually salient stimuli in the CPT. Moreover, in the CPT used here, stimulus-driven orienting has been shown to interact with endogenous attentional processes to facilitate greater vigilance (MacLean et al., 2009). Accordingly, improvements in CPT performance (MacLean et al., 2010) and changes in event-related brain activity may result from improved sensory processing and an interaction between exogenous and endogenous attentional systems (Peterson & Posner, 2012). Exogenous attentional mechanisms may be constrained, however, when improved perceptual capacity cannot be leveraged, as we saw in Retreat 1, when targets were made more difficult to discriminate as discrimination capacity improved.

Rare and difficult-to-detect CPT targets were differentiated from nontargets beginning at the latency of the N1. These findings concur with prior accounts (e.g., Vogel & Luck, 2000), which suggest that modulations in sensory responses occurring at this latency reflect a sensory gain control mechanism involved in selecting task-relevant information for subsequent processing. Indeed, when target difficulty was reparameterized in Retreat 1, individuals who made more difficult discriminations relative to their peers exhibited larger GFP amplitudes at the N1 latency for targets and nontargets, implying that these participants incurred greater processing resources when correctly detecting stimuli. Importantly, the contextual influence of discrimination difficulty on brain responses to nontargets suggests that activity in this epoch, from 150 to 220 msec poststimulus onset, may reflect an early discrimination process that is not purely stimulus-driven but that is also sensitive to top–down influences of attentional demand (Fedota et al., 2012). Finally, we note that, for Retreat 1 training participants only, target stimulus length was positively correlated with GFP amplitude in a window that overlapped with group differences in electric field topography. This suggests that the processing consequences of discrimination difficulty continued into later epochs (i.e., 250–260 msec), which were characterized by group differences in configurations of underlying neural generators.

Two design limitations warrant major consideration. First, the lack of strong experimental control over target difficulty between retreats and, second, the possibility that factors other than the meditation intervention itself, including demand characteristics or effects of repeated task practice, may have contributed to changes across assessments. To more firmly develop mechanistic models of sustained attention training, future studies will need to manipulate task difficulty both between and within subjects in a fully crossed design and extend their investigations to different training intensities, durations, and levels of practitioner experience (King et al., 2019). Though we focus here on data-driven, multivariate methods for assessing global modulations in the neuroelectric field, future investigators might identify subtler effects by using a priori targeted analyses, examining other stimulus conditions (i.e., misses) or isolating cognitive processes through condition subtraction (i.e., difference waves). We hope that the data offered here will aid in the design, interpretation, and evaluation of such studies.

Although behavioral improvements in perceptual discrimination were unambiguous following both training interventions, the consequences of increased perceptual ability on the stimulus processing stream were not. Perceptual improvements might imply plasticity in bottom–up visual systems, but enhancement of sensory signals at early perceptual stages in the CPT appear to result from an attenuation of vigilance-related degradation, perhaps reflecting increased top–down attentional engagement following meditation training. It is clear, however, that the features of exogenous sensory stimuli have a pervasive influence on perceptual and attentional processing during sustained attention. Though further work is needed, our findings help clarify the interactive effects of bottom–up sensory and top–down attentional influences on the stimulus processing stream and their sensitivity to modification through focused-attention meditation.

Acknowledgments

We thank Stephen Aichele, Tonya Jacobs, and David Bridwell for their help in collecting these data and B. Alan Wallace for his many contributions to this study. Major support was provided by Fetzer Institute grant 2191 and John Templeton Foundation grant 39970; the Santa Barbara Institute for Consciousness Studies; and gifts from the Hershey Family, the Baumann, Tan Teo, Yoga Science, and Mental Insight Foundations, and anonymous and other donors all to C. D. S.

Reprint requests should be sent to Anthony Zanesco, Department of Psychology, University of Miami, 5665 Ponce de Leon Blvd, Coral Gables, FL 33146, or via e-mail: apz13@miami.edu.

Notes

1. 

One wait-list control participant withdrew prior to participation in Retreat 2 for reasons unrelated to the intervention.

2. 

The meditation teacher contributed principally to the design of the curriculum and the appropriateness of task measures but was not involved in data collection or analysis.

3. 

Additional paired comparisons between wait-list participants' Retreat 1 and Retreat 2 discrimination threshold revealed that threshold was unchanged from Retreat 1 mid- and postassessment to Retreat 2 preassessment (all ps > .483), but threshold was lower in Retreat 1 compared with Retreat 2 mid- or postassessment (all ps < .009).

4. 

PEST discrimination threshold at preassessment did not predict GFP amplitude in hits or correct rejections at any sustained epoch (no ps < .05 for > 20 msec).

REFERENCES

REFERENCES
Atchley
,
R.
,
Klee
,
D.
,
Memmott
,
T.
,
Goodrich
,
E.
,
Wahbeh
,
H.
, &
Oken
,
B.
(
2016
).
Event-related potential correlates of mindfulness meditation competence
.
Neuroscience
,
320
,
83
92
.
Baumgartner
,
H. M.
,
Graulty
,
C. J.
,
Hillyard
,
S. A.
, &
Pitts
,
M. A.
(
2018
).
Does spatial attention modulate the earliest component of the visual evoked potential?
Cognitive Neuroscience
,
9
,
4
19
.
Belouchrani
,
A.
,
Abed-Meriam
,
K.
,
Cardoso
,
J.-F.
, &
Moulines
,
E.
(
1997
).
A blind source separation technique using second-order statistics
.
IEEE Transactions on Signal Processing
,
45
,
434
444
.
Benjamini
,
Y.
, &
Hochberg
,
Y.
(
1995
).
Controlling the false discovery rate: A practical and powerful approach to multiple testing
.
Journal of the Royal Statistical Society
,
57
,
289
300
.
Boksem
,
M. A.
,
Meijman
,
T. F.
, &
Lorist
,
M. M.
(
2005
).
Effects of mental fatigue on attention: An ERP study
.
Cognitive Brain Research
,
25
,
107
116
.
Brunet
,
D.
,
Murray
,
M. M.
, &
Michel
,
C. M.
(
2011
).
Spatiotemporal analysis of multichannel EEG: CARTOOL
.
Computational Intelligence and Neuroscience
,
2011
,
813870
.
Cahn
,
B. R.
, &
Polich
,
J.
(
2009
).
Meditation (Vipassana) and the P3a event-related brain potential
.
International Journal of Psychophysiology
,
72
,
51
60
.
Dahl
,
C. J.
,
Lutz
,
A.
, &
Davidson
,
R. J.
(
2015
).
Reconstructing and deconstructing the self: Cognitive mechanisms in meditation practice
.
Trends in Cognitive Science
,
19
,
515
523
.
Delgado-Pastor
,
L. C.
,
Perakakis
,
P.
,
Subramanya
,
P.
,
Telles
,
S.
, &
Vila
,
J.
(
2013
).
Mindfulness (Vipassana) meditation: Effects on P3b event-related potential and heart rate variability
.
International Journal of Psychophysiology
,
90
,
207
214
.
Fedota
,
J. R.
,
McDonald
,
C. G.
,
Roberts
,
D. M.
, &
Parasuraman
,
R.
(
2012
).
Contextual task difficulty modulates stimulus discrimination: Electrophysiological evidence for interaction between sensory and executive processes
.
Psychophysiology
,
49
,
1384
1393
.
Ferrer
,
E.
, &
McArdle
,
J. J.
(
2010
).
Longitudinal modeling of developmental changes in psychological research
.
Current Directions in Psychological Science
,
19
,
149
154
.
Folstein
,
J. R.
, &
Van Petten
,
C.
(
2008
).
Influence of cognitive control and mismatch on the N2 component of the ERP: A review
.
Psychophysiology
,
45
,
152
170
.
Guthrie
,
D.
, &
Buchwald
,
J. S.
(
1991
).
Significance testing of difference potentials
.
Psychophysiology
,
28
,
240
244
.
Haubert
,
A.
,
Walsh
,
M.
,
Boyd
,
R.
,
Morris
,
M.
,
Wiedbusch
,
M.
,
Krusmark
,
M.
, et al
(
2018
).
Relationship of event-related potentials to the vigilance decrement
.
Frontiers in Psychology
,
9
,
237
.
Hopf
,
J. M.
,
Vogel
,
E.
,
Woodman
,
G.
,
Heinze
,
H. J.
, &
Luck
,
S. J.
(
2002
).
Localizing visual discrimination processes in time and space
.
Journal of Neurophysiology
,
88
,
2088
2095
.
Jo
,
H. G.
,
Schmidt
,
S.
,
Inacker
,
E.
,
Markowiak
,
M.
, &
Hinterberger
,
T.
(
2016
).
Meditation and attention: A controlled study on long-term meditators in behavioral performance and event-related potentials of attentional control
.
International Journal of Psychophysiology
,
99
,
33
39
.
Kaplan
,
S.
, &
Berman
,
M. G.
(
2010
).
Directed attention as a common resource for executive functioning and self-regulation
.
Perspectives on Psychological Science
,
5
,
43
57
.
King
,
B. G.
,
Conklin
,
Q. A.
,
Zanesco
,
A. P.
, &
Saron
,
C. D.
(
2019
).
Residential meditation retreats: Their role in contemplative practice and significance for psychological research
.
Current Opinion in Psychology
,
28
,
238
244
.
Koenig
,
T.
,
Kottlow
,
M.
,
Stein
,
M.
, &
Melie-García
,
L.
(
2011
).
Ragu: A free tool for the analysis of EEG and MEG event-related scalp field data using global randomization statistics
.
Computational Intelligence and Neuroscience
,
2011
,
14
.
Kok
,
A.
(
2001
).
On the utility of P3 amplitude as a measure of processing capacity
.
Psychophysiology
,
38
,
557
577
.
Luck
,
S. J.
, &
Gaspelin
,
N.
(
2017
).
How to get statistically significant effects in any ERP experiment (and why you shouldn't)
.
Psychophysiology
,
54
,
146
157
.
Lutz
,
A.
,
Slagter
,
H. A.
,
Rawlings
,
N. B.
,
Francis
,
A. D.
,
Greischar
,
L. L.
, &
Davidson
,
R. J.
(
2009
).
Mental training enhances attentional stability: Neural and behavioral evidence
.
Journal of Neuroscience
,
29
,
13418
13427
.
MacLean
,
K. A.
,
Aichele
,
S. R.
,
Bridwell
,
D. A.
,
Mangun
,
G. R.
,
Wojciulik
,
E.
, &
Saron
,
C. D.
(
2009
).
Interactions between endogenous and exogenous attention during vigilance
.
Attention, Perception, & Psychophysics
,
71
,
1042
1058
.
MacLean
,
K. A.
,
Ferrer
,
E.
,
Aichele
,
S. R.
,
Bridwell
,
D. A.
,
Zanesco
,
A. P.
,
Jacobs
,
T. L.
, et al
(
2010
).
Intensive meditation training improves perceptual discrimination and sustained attention
.
Psychological Science
,
21
,
829
839
.
Murray
,
M. M.
,
Brunet
,
D.
, &
Michel
,
C. M.
(
2008
).
Topographic ERP analyses: A step-by-step tutorial review
.
Brain Topography
,
20
,
249
264
.
Nuechterlein
,
K. H.
,
Parasuraman
,
R.
, &
Jiang
,
Q.
(
1983
).
Visual sustained attention: Image degradation produces rapid sensitivity decrement over time
.
Science
,
220
,
327
329
.
Parasuraman
,
R.
(
1979
).
Memory load and event rate control sensitivity decrements in sustained attention
.
Science
,
205
,
924
927
.
Parasuraman
,
R.
,
Warm
,
J. S.
, &
See
,
J. E.
(
1998
).
Brain systems of vigilance
. In
R.
Parasuraman
(Ed.),
The attentive brain
.
Cambridge, MA
:
MIT Press
.
Perrin
,
F.
,
Pernier
,
J.
,
Bertrand
,
O.
, &
Echallier
,
J. F.
(
1989
).
Spherical splines for scalp potential and current density mapping
.
Electroencephalography and Clinical Neurophysiology
,
72
,
184
187
.
Peterson
,
S. E.
, &
Posner
,
M. I.
(
2012
).
The attention system of the human brain: 20 Years after
.
Annual Review of Neuroscience
,
35
,
73
89
.
Polich
,
J.
(
2007
).
Updating P300: An integrative theory of P3a and P3b
.
Clinical Neurophysiology
,
118
,
2128
2148
.
Saggar
,
M.
,
King
,
B. G.
,
Zanesco
,
A. P.
,
MacLean
,
K. A.
,
Aichele
,
S. R.
,
Jacobs
,
T. L.
, et al
(
2012
).
Intensive training induces longitudinal changes in meditation state-related EEG oscillatory activity
.
Frontiers in Human Neuroscience
,
6
,
256
.
Skrandies
,
W.
(
1990
).
Global field power and topographic similarity
.
Brain Topography
,
3
,
137
141
.
Slagter
,
H. A.
,
Davidson
,
R. J.
, &
Lutz
,
A.
(
2011
).
Mental training as a tool in the neuroscientific study of brain and cognitive plasticity
.
Frontiers in Human Neuroscience
,
5
,
17
.
Taylor
,
M. M.
, &
Creelman
,
C. D.
(
1967
).
PEST: Efficient estimates on probability functions
.
Journal of the Acoustical Society of America
,
41
,
782
787
.
van Leeuwen
,
S.
,
Singer
,
W.
, &
Melloni
,
L.
(
2012
).
Meditation increases the depth of information processing and improves the allocation of attention in space
.
Frontiers in Human Neuroscience
,
6
,
133
.
Vaughan
,
H. G.
(
1982
).
The neural origins of human event-related potentials
.
Annals of the New York Academy of Sciences
,
388
,
125
138
.
Vogel
,
E. K.
, &
Luck
,
S. J.
(
2000
).
The visual N1 component as an index of a discrimination process
.
Psychophysiology
,
37
,
190
203
.
Wadlinger
,
H. A.
, &
Isaacowitz
,
D. M.
(
2011
).
Fixing our focus: Training attention to regulate emotion
.
Personality and Social Psychology Review
,
15
,
75
102
.
Wallace
,
B. A.
(
1999
).
The Buddhist tradition of Samatha: Methods for refining and examining consciousness
.
Journal of Consciousness Studies
,
6
,
175
187
.
Wallace
,
B. A.
(
2006
).
The attention revolution
.
Boston, MA
:
Wisdom Publications
.
Zanesco
,
A. P.
,
King
,
B. G.
,
MacLean
,
K. A.
, &
Saron
,
C. D.
(
2013
).
Executive control and felt concentrative engagement following intensive meditation training
.
Frontiers in Human Neuroscience
,
7
,
566
.
Zanesco
,
A. P.
,
King
,
B. G.
,
MacLean
,
K. A.
, &
Saron
,
C. D.
(
2018
).
Cognitive aging and long-term maintenance of attentional improvements following meditation training
.
Journal of Cognitive Enhancement
,
2
,
259
275
.