Brain signal variability holds valuable information about brain dynamics that should not be ignored when studying the link between neuronal activity and cognition. Brain signal variability can take many forms. Moment-to-moment variability of the ongoing signal reflects how much brain activity changes from one moment to the next, that is the brain activity range. Trial-by-trial variability in task-related evoked responses reflects differences in the brain responses to repeated task conditions, while brain signal entropy quantifies the irregularity of the time series.

Within-subject brain signal variability has been shown to change with ageing and is a robust marker of age (Grady and Garrett, 2014). The effect of ageing on brain signal variability has been studied using several measures, including temporal standard deviation (SD) and mean square successive difference (MSSD) of the blood-oxygen-level-dependent (BOLD) signal acquired with functional magnetic resonance imaging (fMRI) (Grady and Garrett, 2014; Samanez-Larkin et al., 2010), temporal SD of the spectral power within frequency bands measured with electroencephalography (EEG) (Kumral et al., 2020), entropy in EEG and magnetoencephalography (MEG) signals (Kosciessa et al., 2020; McIntosh et al., 2014; Waschke et al., 2017), the slope of the EEG power spectral density (PSD) (Voytek et al., 2015), and the consistency of task-related evoked responses (Sander et al., 2012; Tran et al., 2020). The temporal standard deviation of the BOLD signal and of the amplitude envelop of EEG frequency bands measured in the ongoing signal are markedly decreased in older adults (Garrett et al., 2015; Grady and Garrett, 2018; Kumral et al., 2020). Yet, the localization of these changes does not coincide suggesting that BOLD and EEG variability capture different aspects of brain function. In fact, age differences in BOLD SD might be explained by changes in cardiovascular and cerebrovascular factors that affect the link between neuronal activity and changes in blood flow (Tsvetanov et al., 2021). Moreover, BOLD MSSD, a measure of how much the BOLD signal changes across successive observations, is more often increased rather than decreased in older adults (Boylan et al., 2021; Nomi et al., 2017; Samanez-Larkin et al., 2010), highlighting the fact that age-related changes in signal dynamics can result in more or less ‘signal variability’ depending on the used metric. In fact, the relationship between the non-invasive measures of brain signal variability used in human neuroscience and neuronal noise is not trivial. Large signal fluctuations will increase signal standard deviation but might originate in a more stable and predictable signal associated with reduced noise. Higher signal variability is, therefore, not equal to a noisier signal. Consistent with this idea, EEG signal entropy increases with ageing and is associated with the EEG spectra slope (Waschke et al., 2017). Older adults present EEG signals with flatter spectra, i.e., more similar to white noise, potentially reflecting increased neuronal noise (Voytek et al., 2015). Higher brain noise is also consistent with the fact that older people present increased behavioral variability measured as less consistent reaction time across trials (Grady and Garrett, 2018; MacDonald et al., 2012). However, the link between brain signal variability, neuronal noise and behavioral variability remains to be clarified.

The ongoing brain signal (i.e. spontaneous brain activity) measured with fMRI or EEG during ‘task-free’, ‘resting’ periods reflects activity in large-scale networks each identified as a set of brain regions that show activity co-fluctuations, the resting-state networks (Beckmann et al., 2005; Liu et al., 2017). Notably, these networks are functionally active also during task performance (Abreu et al., 2020; Smith et al., 2009). The ongoing activation and de-activation of these networks presents a backdrop of activity on top of which task-related activity occurs (Fox et al., 2006). Not surprisingly, ongoing activity modulates task-related evoked responses that in turn affect task performance (Becker et al., 2011; Fox et al., 2007; Mayhew et al., 2013; Ribeiro et al., 2016). For example, fluctuations in pre-stimulus alpha power affect the amplitude of brain sensory responses, perception and reaction time (Lou et al., 2014; Thut et al., 2006; van Dijk et al., 2008). Older people display low-frequency (<12 Hz) brain signal fluctuations with reduced amplitude (Kumral et al., 2020), however, previous studies of the effect of ageing on the variability of task evoked responses have yielded contradictory results, revealing increased or decreased trial-by-trial consistency depending on stimulus or task characteristics (Sander et al., 2012; Tran et al., 2016; Tran et al., 2020; Wiegand and Sander, 2019). Finally, to our knowledge, no previous study has investigated the effect of ageing on the link between the dynamics of the ongoing brain activity, trial-by-trial variability of evoked responses and behavioral variability.

In this study, we aimed at clarifying the effect of ageing on brain signal dynamics and its impact on behavior by investigating the relationship between variability of ongoing brain signals, evoked responses, and behavior. We analysed EEG and pupil data from a previously published study acquired while a group of young and a group of older adults were engaged in a cued auditory go/no-go task (Ribeiro and Castelo-Branco, 2019a; Ribeiro and Castelo-Branco, 2019b; Ribeiro and Castelo-Branco, 2021). The EEG allows the non-invasive measurement of cortical electrical activity with high temporal resolution, while the pupillogram, acquired under conditions of constant luminance, is associated with fluctuations in arousal and reflects activity in the ascending brainstem neuromodulatory systems (de Gee et al., 2017; Joshi et al., 2016; Murphy et al., 2014; Reimer et al., 2016). Although, activity modulation in subcortical arousal nuclei (reflected in changes in pupil size) correlates with the EEG signal (e.g. Hong et al., 2014; Podvalny et al., 2021), these two signals capture distinct aspects of brain function, both important and complementary for our understanding of the ageing brain. In our previous study, we showed that, in both signals, the cue stimulus evoked a preparatory response. In the EEG, the cue evoked a frontocentral negative potential [the contingent negative variation (CNV)] that increased in magnitude throughout the preparatory period. The CNV has been previously shown to correlate with reaction time on a trial-by-trial basis (Boehm et al., 2014), suggesting that trial-by-trial amplitude fluctuations in the CNV are associated with behavioral variability. In the pupillogram, the cue evoked a pupil dilation response of equal magnitude in both groups of participants (Ribeiro and Castelo-Branco, 2019a). In the current study, we investigated how trial-by-trial variability in the amplitude of these preparatory responses changed across age groups, how it related to behavior variability and how it was affected by ongoing brain signals. We found that, while behavioral variability was increased in older adults, this group of participants presented reduced variability of the evoked responses both in the EEG and in the pupillogram. We found that the decreased variability in the evoked responses emerged due to a reduction in the amplitude of the slow fluctuations observed in the ongoing signals and did not reflect reduced variability of the evoked responses per se. Once differences in the dynamics of the ongoing signals were accounted for, we observed no group differences in the variability of the evoked responses. Moreover, although the evoked responses were markedly affected by fluctuations in the ongoing signals, the amplitude modulations caused by the ongoing signal did not contribute towards behavioral variability, thereby explaining why although ongoing brain signal variability is decreased in older individuals, behavioral variability is not.

We analysed EEG and pupil data acquired while a group of young (N=36; age = 23 ± 3 years) and a group of older (N=39; age = 60 ± 5 years) adults were engaged in a cued-auditory go/no-go task (Figure 1A).

Figure 1

Download asset Open asset

Variability in the amplitude of preparatory evoked responses is decreased in the older group and is associated with reaction time

Trial-by-trial variability in the CNV amplitude (CNV variability) was decreased in the older group (Figure 2). During the preparatory period between the cue and the target, a frontocentral negative potential (the CNV) could be observed in the EEG of both groups of participants with strongest amplitude in the frontocentral electrode FCz (Figure 2A and C). Although the amplitude of this evoked response was on average stronger (more negative) in the older group, trial-by-trial variability (across trials standard deviation) was reduced in the older group (Figure 2B and D; see Figure 2—figure supplement 1f for single subject examples of trial-by-trial CNV variability). We observed significant group differences in the CNV amplitude and variability (Figure 2E and F). However, the group differences showed different topographic distributions. CNV amplitude presented strongest group difference in left central electrodes, while CNV variability presented strongest group difference in parieto-occipital electrodes.

Figure 2 with 1 supplement see all

Download asset Open asset

Trial-by-trial CNV amplitude fluctuations were associated with reaction time fluctuations (Figure 3). Within-subject correlation analyses revealed a significant relationship between CNV amplitude and reaction time. Although, the correlation coefficients were on average smaller (closer to zero) in the young group, these did not differ significantly across groups (|ts|<2.6). The correlation values were significantly different from zero in a left frontocentral cluster of electrodes (one-sample t-tests including all participants; Figure 3A). As participants were responding with their right index finger the fact that the stronger correlation coefficients were observed in a left central cluster (contralateral to the response hand) suggests that the preparatory activity measured has a motor element. The CNV is comprised of different components including a motor-related late component (Nagai et al., 2004; van Boxtel and Böcker, 2004). This result suggests that the association between CNV amplitude and reaction time results from preparatory motor activity happening during that period.

Figure 3

Download asset Open asset

Older individuals presented reduced trial-by-trial variability in cue-locked pupil dilation (PD) responses (Figure 4). The warning cue elicited a pupil dilation response that peaked around 3 s after cue-onset (as reported before in Ribeiro and Castelo-Branco, 2019a; replotted in Figure 4A). PD response amplitude was not significantly different across groups [t₍₇₁₎=–1.82, p=0.073]. The variability of PD response amplitude was, however, significantly reduced in the older group [t₍₇₁₎=6.05, p<0.001; Figure 4B; see Figure 4—figure supplement 1 for single subject examples of trial-by-trial variability in evoked responses].

Figure 4 with 1 supplement see all

Download asset Open asset

Preparatory pupil dilation responses did not correlate with reaction time. Within-subject correlation coefficients were not significantly different across groups [t₍₇₁₎=1.10, p=0.276] or significantly different from zero [one-sample t-test including all participants, t₍₇₂₎ = 0.670, p=0.505].

In summary, trial-by-trial variability in the amplitude of the evoked responses was markedly reduced in older adults, while reaction time variability was increased. These observations suggest that the additional variability observed in the young group originates in a signal component that is not behaviorally relevant. In the following sections, we will delve into the origin of the variability in these evoked responses.

Spectral properties of the ongoing EEG and pupil signals

Evoked neural responses occur on a background of ongoing neural activity that unavoidably adds to the responses measured (Fox et al., 2006; Shimaoka et al., 2019). It is possible therefore that age-dependent changes in the dynamics of ongoing activity explain the differences in the variability of task-related evoked responses. We explored the relationship between ongoing signal dynamics and the variability of the evoked responses by studying the spectral properties of ongoing EEG and pupil data and how these relate to the evoked responses.

The ongoing EEG signal of older adults presented power spectral density (PSD) data that are less steep, with reduced offset (reduced amplitude at low frequencies), reduced alpha power and increased beta power (Figure 5). Ongoing EEG activity was estimated from the pre-stimulus period (3.5 s) just before cue onset, taking advantage of the long inter-trial intervals used. We analysed the PSDs using the FOOOF toolbox that separates the aperiod from the periodic component of the spectra (Donoghue et al., 2020). The aperiodic component can be described by its two parameters: the spectral exponent (a measure of how steep the spectrum is) and the offset (which reflects the uniform shift of power across frequencies). Both were significantly reduced in the older group (Figure 5B). Occipital alpha power was reduced in older individuals mainly in occipital channels. In contrast, beta power in frontal and central EEG channels was increased in the older group.

Figure 5 with 1 supplement see all

Download asset Open asset

The PSDs of ongoing pupil signals of older individuals presented reduced exponent and offset (Figure 6). Ongoing pupil signal fluctuations happen on a slower time scale than the EEG fluctuations, therefore, it is important to analyse longer time periods to study their dynamics. To analyse ongoing pupil fluctuations in the absence of task related activity, we took advantage of the 4 min long pupillary recordings obtained at the beginning of the acquisition protocol while participants were fixating and passively listening to the cue stimulus being presented with the same frequency as in the task (Ribeiro and Castelo-Branco, 2019a). Besides fixating the centre of the screen no other overt task was required. Due to the absence of task or luminance changes, the dynamics of pupil fluctuations in these recordings can be mostly assigned to ongoing pupillary activity. We analysed the spectral content of these signals in 20 s epochs cut sequentially from these 4 min recordings (not locked to the auditory stimulus).The spectra of the ongoing pupillary signal followed a 1 /f distribution, with power decreasing with increasing frequency (Figure 6). The aperiodic parameters were significantly decreased in the older group [exponent: t₍₇₀₎ = 7.48, p<0.001; offset: t₍₇₀₎ = 4.17, p<0.001], that is the older group presented a flatter pupil spectrum with lower amplitude at the low frequencies. No obvious peaks were detected in the pupil spectra.

Figure 6 with 1 supplement see all

Download asset Open asset

Reduced variability in the evoked responses of older adults is linked to reduced amplitude in the slow fluctuations of ongoing activity

Spectral properties of ongoing signals are associated with trial-by-trial variability in evoked responses (Figure 7). We used correlational analyses to investigate if the dynamics of the ongoing EEG and pupil signals captured by the spectral parameters was associated with the variability in task-related evoked responses across participants. For the EEG, to facilitate data presentation and interpretation of the following analyses, we focused on the FCz channel where the CNV showed its highest amplitude. Similar findings were observed in the other EEG channels, including the ones presenting significant correlations with behavior (Figure 7B and Figure 7—figure supplement 1). Participants with higher spectral offset and exponent presented higher CNV and PD variability (Figure 7). Alpha power presented a positive association with CNV variability only in the young group and in posterior EEG channels (Figure 7B). Beta power was not associated with CNV variability. These findings suggest that the aperiodic signal fluctuations are intrinsically linked to the variability observed in the evoked responses. We then sought to determine which frequencies in the aperiodic spectra were more predictive of the variability in the amplitude of the evoked responses. We found that the correlation between variability in the amplitude of the evoked responses and the amplitude of the power spectra was particularly strong in the lower frequencies (Figure 7—figure supplement 1). Importantly, the correlations were significant in both the aperiodic component of the fitted (FOOOF) spectra and in the raw (total) power spectra. In the total spectral power, the correlation between PD variability and spectral power was highest around 0.44 Hz and decreased for lower frequencies. In the EEG analyses, correlation was highest at 1 Hz, the lowest frequency analysed. These findings suggest that slow aperiodic fluctuations observed in the ongoing signals are associated with trial-by-trial variability in the amplitude of the evoked responses, both in the EEG and in the pupillogram.

Figure 7 with 1 supplement see all

Download asset Open asset

The group differences in CNV and PD variability were markedly reduced once the effect of the spectral exponent, offset or slow spectral power were regressed out (Figure 8). This was evident in the significant interaction effects observed in repeated measures ANOVAs including as within-subject factor adjustment for confounds (with and without adjustment) and group as between-subject factor [CNV variability adjustment x group interaction effect when adjusting for exponent F_{(1, 72)}=47.2, p<0.001, offset F_{(1, 72)}=14.3, p<0.001, or slow spectral power F_{(1, 70)}=13.4, p<0.001; PD variability adjustment x group interaction effect when adjusting for exponent F_{(1, 71)}=54, p<0.001, offset F_{(1, 71)}=19, p<0.001, or slow spectral power F_{(1, 71)}=62, p<0.001]. In fact, after adjusting for the confounds, the CNV variability was no longer significantly different across groups [adjusting for exponent t₍₇₂₎ = 0.661, p=0.511; adjusting for offset t₍₇₂₎ = 0.243, p=0.809; adjusting for slow spectral power t₍₇₂₎ = 0.979, p=0.331; Figure 8]. Similar findings were observed in the other EEG channels after controlling for multiple comparisons. After adjusting for the exponent, group differences in CNV variability were observed only in two posterior EEG channels (PO8 and CB2). No channels showed group differences after adjusting for offset or slow spectral power. The group difference in the PD variability was also completely explained by the power of slow fluctuations observed in the ongoing pupillary signal [t-test group comparison no longer significant: t₍₇₁₎ = 1.54, p=0.129; Figure 8]. However, adjusting for spectral exponent or offset did not completely explain the group difference [adjusting for exponent t₍₇₁₎ = 2.45, p=0.017; adjusting for offset t₍₇₁₎ = 3.65, p<0.001]. These findings indicate that the age-related reduction in the power of the slow aperiodic fluctuations underlie the group differences in trial-by-trial variability in the amplitude of the evoked responses.

Figure 8

Download asset Open asset

Factors that contribute to trial-by-trial variability in evoked responses

There are two possible mechanisms through which background signal fluctuations affect the amplitude of the evoked responses. Background signal fluctuations might reflect fluctuations in brain state and different brain states will be associated with different evoked responses; and/or background signal fluctuations and the evoked responses are two independent signals that sum giving rise to a composite signal where the evoked response appears on top of a fluctuating baseline thereby modulating its shape and amplitude - a simple additive mechanism. In this latter case, the phase of the background fluctuations at the cue-onset rather than the instantaneous amplitude would better capture the relationship between the measured evoked responses and the ongoing signal (at the same amplitude values, the phase of the oscillation will define if the signal is increasing or decreasing). We explored this relationship by estimating the phase of the slow fluctuations in the EEG and pupil signals at cue-onset. For illustrative purposes, we plotted the relationship between the pre-stimulus phase and the amplitude of the respective evoked responses (Figure 9). The evoked responses showed a dependence on the phase of the ongoing fluctuations in line with what would be expected if the measured evoked responses were the summation of an evoked response with an ongoing fluctuating signal (Figure 9B and C). This relationship was also different from what would be expected if the evoked responses were affected solely by the amplitude of the baseline signal. This is further explored quantitatively in the next paragraph.

Figure 9

Download asset Open asset

Ongoing signal fluctuations strongly affected the amplitude of the evoked responses (Figure 10). We quantified the association between EEG/pupil phase of slow fluctuations estimated at cue-onset (θ) and the amplitude of the respective evoked responses using linear regression. This relationship can be captured in a linear regression by representing the phase as a pair of variables: the cosine and the sine (Nurhab et al., 2017). We multiplied the cosine and sine functions by the amplitude envelop (r) of the slow signal fluctuations because the amplitude envelop varies throughout the recordings and these fluctuations will have differential effects on the evoked responses depending on their amplitude. The two predictors associated with the slow ongoing signal fluctuations were r • sin θ and r • cos θ. Note that r • cos θ is equal to the signal amplitude and therefore to the amplitude at baseline, while r • sin θ will capture the additional effect of the signal phase. To study how these factors affected the trial-by-trial amplitude of the task-related responses, we run generalized linear regression models (GLMs) for each participant and separately for each signal type - EEG and pupil. We included as dependent variable the amplitude of the evoked response and as predictors r • sin θ, r • cos θ and, additionally, time-on-task (that can lead to differential responses due to learning, adaptation, or increased fatigue), and pre-stimulus alpha power (that reflects moment-to-moment spontaneous fluctuations in brain state related to alertness or attentional fluctuations; van Diepen et al., 2015). Run was included as categorical predictor (two consecutive runs of 8 min each were acquired). These models explained a large amount of the evoked responses variance (R² mean ± SE: EEG young 67% ± 8%; EEG older 63% ± 8%; pupil young 73% ± 1%; pupil older 67% ± 1%). The explained variances were significantly decreased in the alternative model without the r • sin θ predictor that captured the effect of the ongoing signal phase [R² mean ± SE without r • sin θ: EEG young 27% ± 11%; EEG older 28% ± 10%; pupil young 40% ± 1%; pupil older 37% ± 1%, paired t-test comparing the two models’ explained variance: EEG t₍₇₃₎ = –26.8, p<0.001; pupil t₍₇₁₎ = –22.9, p<0.001], suggesting that the effect of the ongoing signal on the amplitude of the evoked responses is best captured by including amplitude and phase of the ongoing signal at baseline. We tested the estimated coefficients against zero using one-sample t-tests. The EEG GLM revealed that the CNV amplitude depended on time-on-task [it became less negative with time-on-task; t₍₇₃₎ = 2.84, p=0.006; Figure 9A], and on the phase of the ongoing fluctuations at cue-onset captured by the cosine [t₍₇₃₎=–36.3, p<0.001] and the sine [t₍₇₃₎=–81.7, p<0.001] functions (Figure 10A). It was independent of run [t₍₇₃₎=0.035, p=0.973] and pre-stimulus alpha power [t₍₇₃₎=–0.530, p=0.598]. Time-on-task had a stronger effect on the evoked responses of the young group in comparison with the older group [independent samples t-test: t₍₇₂₎ = 3.74, p<0.001]. The other coefficients did not present significant effects of group [run: t₍₇₂₎ = 1.98, p=0.051; pre-stimulus alpha: t₍₇₂₎ = –0.098, p=0.923; cosine: t₍₇₂₎ = 0.189, p=0.851; sine: t₍₇₂₎ = –0.007, p=0.994]. The pupil GLM revealed that the only predictors that significantly contributed towards trial-by-trial PD variability were the cosine and the sine of the pre-stimulus phase [one-sample t-test testing estimated coefficients against zero - cosine: t₍₇₁₎ = –38.2, p<0.001; sine: t₍₇₁₎ = –51.3, p<0.001]. The other predictors were not significantly different from zero [run: t₍₇₁₎ = 1.88, p=0.064; time-on-task: t₍₇₁₎ = –1.34, p=0.185; alpha power: t₍₇₁₎ = –0.061, p=0.952]. None of the coefficients showed a significant group difference [run: t₍₇₀₎ = –0.198, p=0.844; time-on-task: t₍₇₁₎ = –1.00, p=0.321; alpha power: t₍₇₀₎ = 1.68, p=0.098; cosine: t₍₇₀₎ = –0.143, p=0.887; sine: t₍₇₀₎ = –0.396, p=0.693]. These results suggest a strong effect of the amplitude and phase of the ongoing signal fluctuations on the evoked responses (Figure 10A and B), consistent with the idea that there is a drift in the signal baseline that adds to the evoked responses pushing them towards positive or negative values depending on the phase at cue-onset.

Figure 10

Download asset Open asset

To test if the variability in the evoked responses associated with the ongoing signal fluctuations was behaviorally relevant, we studied how adjusting for the signal phase at baseline changed the correlation between the evoked responses and reaction time. After regressing out the effect of the ongoing signal fluctuations from the amplitude of the evoked responses, these were associated with reaction time at least as strongly as without the adjustment. In particular, the CNV amplitude residuals correlated with reaction time in a larger number of electrodes than without the adjustment and unmasked a significant relationship between the CNV amplitude and reaction time in frontal electrodes that was not observed without the adjustment (Figure 11A), suggesting that the relationship between CNV amplitude and behavior is independent of the baseline fluctuations. On average the correlation coefficients with the adjustment were slightly more positive than without the adjustment. This difference was significantly different in three EEG channels (F4, C1, and C4) after correcting for multiple comparisons. For the pupil responses, we found that regressing out the effect of ongoing signal fluctuations was crucial to detect the link between PD evoked responses and reaction time, that is the PD amplitude residuals correlated with reaction time [Figure 11C; one-sample t-test: t₍₇₂₎ = –4.77, p<0.001]. Notably, the correlation coefficients with the adjustment for the phase of the ongoing signal were significantly different from the coefficients without the adjustment [paired t-test comparing with and without adjustment: t₍₇₂₎ = –5.08, p<0.001]. These findings suggest that the effect of the ongoing signal on the evoked responses is not behaviorally relevant and can mask brain-behavior associations.

Figure 11 with 1 supplement see all

Download asset Open asset

It is important to note that after regressing out the effect of the ongoing signal’s phase at cue-onset, the variability of the evoked responses was still significantly reduced in the older group (Figure 11—figure supplement 1), suggesting that the baseline (pre-stimulus) values are not enough to completely predict the effect of the ongoing signal fluctuations on the evoked responses in line with the idea that these slow signal fluctuations are aperiodic and therefore not completely predictable (Shimaoka et al., 2019).

Joint contribution of pupil dilation and CNV responses to reaction time variability

After adjusting the evoked responses for the effect of the ongoing slow signal fluctuations, both the pupil dilation and the CNV responses correlated with reaction time. However, it was not clear if and how their contributions to the behavioral variability interacted. To address this question, we first checked if these two responses correlated with each other. The pupil dilation response shows a delay of hundreds of milliseconds in relation to the neural responses with which it is associated (Joshi et al., 2016; Joshi and Gold, 2022; Reimer et al., 2016; Yang et al., 2021). To take into account the delay in the pupil response, we correlated the CNV amplitude measured in the time window at the end of the preparatory period (1–1.5 s after cue onset) with the pupil dilation response measured in three time windows: one concomitant with the CNV amplitude measurements as used in the previous analyses (1–1.5 s after cue onset), and two later time windows (1.5–2 s and 2–2.5 s after cue onset). In within-subject correlation analyses, we found that the single trial measurements of the CNV and pupil responses were significantly correlated with stronger correlation coefficients observed in posterior channels (Figure 12A and B). The correlation coefficients were stronger when the pupil response was measured 500ms after the CNV measurements. Just like for the correlations with task performance, these correlations were stronger after adjusting for the ongoing slow signal fluctuations. The effect of pupil time window was stronger in the older group and the correlation coefficients showed a stronger effect of adjustment for the later time windows. Statistically, in a repeated measures ANOVA including as within-subject factors the effect of adjustment and pupil time window, and group as between-subjects factor, across channels average correlation coefficients presented a significant effect of pupil time window [F_{(2, 70)}=9.43, p<0.001 ], a significant effect of adjustment [F_{(1, 70)}=4.59, p=0.036 ], a significant interaction pupil time window x group [F_{(2, 70)}=3.26, p=0.041], and a significant interaction pupil time window x adjustment [F_{(2, 140)}=7.890, p<0.001]. The across channels average correlation coefficients showed no effect of group [F_{(1, 70)}=1.17, p=0.283], and no group x adjustment interaction [F_{(1, 70)}=0.661, p=0.419] and no three-way interaction pupil time window x adjustment x group [F_{(2, 140)}=2.43, p=0.092].

Figure 12

Download asset Open asset

As these evoked responses were correlated, it is possible that one of them is mediating the effect of the other on reaction time. We tested this using multiple regression models with reaction time as dependent variable and CNV and/or PD residuals adjusted for the signal phase at baseline as independent variables. We were interested in understanding the cause of the behavioral variability observed. Therefore, we focused on the neural responses that occur before the button press and only included in these analyses the pupil dilation response measured during the preparatory period before target onset (measured at the same time window as the CNV). The reaction time variance explained by the model including CNV and pupil residuals was significantly higher than the variance explained by the model including only the pupil dilation residuals or the model including only the CNV residuals [Figure 12C; R² across participants mean ± SD: pupil only model young group = 2% ± 3%; pupil only model older group = 2% ± 3%; CNV only model young group (average across channels)=2% ± 2%; CNV only model older group (average across channels)=4% ± 4%; pupil +CNV model young group (average across channels)=4% ± 3%; pupil +CNV model older group (average across channels)=7% ± 4 %]. In repeated measures ANOVA with model as within-subject factor (three models: pupil only, CNV only, and pupil and CNV together) and group as between-subject factor, the across channels average correlation coefficients showed a significant effect of model [F_{(2, 140)}=33.5, p<0.001], no group x model interaction [F_{(2, 140)}=3.06, p=0.050] and a significant group effect [F_{(1, 70)}=6.14, p=0.016] with the older group presenting higher R² values. Post hoc tests revealed that the CNV and pupil model had significantly higher R² in comparison with the pupil only model and the CNV only model (ps <0.001), while the reaction time variances explained by the pupil only and CNV only models were not significantly different (p=0.056). These findings suggest that the pupil dilation response and the CNV contribute independently towards behavior variability.

Task-related brain evoked responses occur on top of ongoing signal fluctuations that markedly affect their amplitude on a trial-by-trial basis. Our results revealed that as a consequence of older people showing reduced amplitude in the ongoing slow signal fluctuations measured in the EEG and pupil signals, trial-by-trial variability in their evoked responses is also decreased. Once the difference in ongoing signal dynamics was adjusted for, the evoked responses were equally variable in young and older adults. This finding is consistent with the fact that behavioral variability does not decrease with age as would be expected if the evoked responses were more consistent across trials. In fact, although ongoing signal fluctuations strongly affected the amplitude of the evoked responses, these amplitude modulations did not appear to have behavioral consequences.

There are two possible ways through which the slow fluctuations observed in the ongoing signals impact the evoked responses. One possibility is that these slow signal fluctuations reflect fluctuations in brain state that affect the way the brain responds to external stimuli as has been suggested for example for the impact of infra-slow oscillations observed in EEG (Monto et al., 2008) or pre-stimulus activity levels in resting-state networks in fMRI (Boly et al., 2007). The other possibility assumes that the effect of the ongoing fluctuations on the evoked responses is purely additive and simply reflects the fluctuating baseline and is therefore not relevant for the type of task studied. This second hypothesis is supported by previous fMRI studies in humans (Fox et al., 2006; Schölvinck et al., 2012) as well as studies of neuronal responses measured in the sensory cortices of mice and cats (Schölvinck et al., 2015; Shimaoka et al., 2019). Our results further suggest that this latter case dominates the effect of the ongoing EEG and pupil signals on their respective evoked responses, highlighting an effect of ongoing signal’s phase: if the ongoing signal is increasing in amplitude, then the evoked response shows a more positive amplitude and vice-versa. Indeed, similar baseline amplitude levels at different signal phases (with amplitude increasing or amplitude decreasing) have different effects on the amplitude of the evoked responses (see Figure 9) and the variance of the evoked responses explained by the models tested increased substantially by including the ongoing signal phase [r • sin(θ)] at cue-onset in comparison with including only the signal amplitude (as captured by the cosine function). Importantly, when we adjusted for the effect of the ongoing fluctuations on the evoked response’s amplitude, the correlation between the evoked response and reaction time was at least as strong as without this adjustment, suggesting that the effect of the ongoing signal fluctuations on the evoked responses has no impact on behavior. Interestingly, in both the pupil and EEG signals adjusting for ongoing signal fluctuations unmasked a relationship between the evoked responses amplitude and reaction time that, without the adjustment, was not detectable in the pupil and was evident only in a restricted number of channels in the EEG. This observation suggests that the additional variability imposed on the evoked responses by the ongoing signal can mask brain-behavior associations and therefore should be attended to and adjusted for.

Although, the effect of the ongoing signal on the amplitude of the evoked responses was not behaviorally relevant, ongoing signal fluctuations might reflect changes in brain activity that affect task performance independently from their impact on the evoked responses. In fact, Schölvinck et al., 2012, using fMRI data, found that spontaneous activity was associated with visual perception without significantly modulating visual evoked responses. Moreover, an effect of baseline (pre-stimulus) slow activity fluctuations on task performance has been observed in studies using EEG or pupil data (Gilzenrat et al., 2010; Monto et al., 2008). Accordingly, we explored the relationship between pre-stimulus slow brain signal fluctuations and reaction time, however, it is important to note that as we were using a cued task our pre-stimulus activity refers to activity before cue onset that is at least 1.5 s before target onset. Under these conditions, pre-cue slow signal fluctuations in the EEG were associated with reaction time in two left temporal channels. We found that in these EEG channels the sine of the phase of the ongoing signal was negatively related to reaction time suggesting that faster reactions happen during the descending phase of the ongoing slow EEG fluctuations. In the pupil data, we observed a significant negative relationship between pupil size (captured by the cosine of the signal phase) and reaction time suggesting that larger pupils at cue onset are associated with faster reactions. Large pupils might reflect higher levels of arousal and therefore better task engagement. In fact, this is also what we observe in our pupil evoked response – the larger the pupil response the faster the reaction times. The elucidation of the mechanisms by which baseline pupil size affects task performance is hampered by the fact that baseline pupil is intrinsically linked to the phase of the ongoing slow fluctuations that in turn affect the amplitude of the evoked responses in an additive manner and, in this case, in the opposite direction from the one predicted from its effect on behavior: large baseline pupils are associated with fast reaction times but are also associated with smaller pupil evoked responses that are in turn associated with slower reaction times. This observation further supports the idea that the effect of pupil baseline on behavior is independent of the effect of the pupil baseline on the evoked responses.

The association between baseline signal fluctuations and reaction time was not significantly different across groups suggesting that signal fluctuations affect reaction time similarly in young and older people. Larger signal fluctuations in the young group would suggest higher reaction time variability. This is not the case, as young participants display lower reaction time variability. It is therefore unlikely that this mechanism contributes towards the age-related differences in behavioral variability.

Fluctuations in pupil size under constant illumination conditions reflect activity in a restricted number of brain regions. The main areas that have been identified in non-luminance modulation of pupil size are the superior colliculus and the locus coeruleus (LC; Joshi and Gold, 2020). The involvement of the superior colliculus in pupil dilation responses is thought to predominate during the orienting response to sensory stimuli (Wang et al., 2014), while activation of the LC modulates pupil size during resting periods as well as during cognitive arousal (Joshi et al., 2016; Murphy et al., 2014; Reimer et al., 2016). However, the pupil inverse problem (which brain areas are being modulated when we observe changes in pupil size) remains to be solved (Joshi and Gold, 2020). Our results suggest that the fluctuations observed in the ongoing pupil signal and the pupil evoked responses correlate with reaction time even though the effect of the ongoing signal on the evoked responses was not behaviorally relevant. The component of the ongoing pupil signal associated with task performance might be separate from the large fluctuations observed and might even correspond to the same mechanism that is further activated in response to the cue. It is likely that the component of the ongoing pupil signal that is behaviorally relevant constitutes a small part of the signal. In fact, studies describing the association between neural activity and ongoing pupil dilation events report significant but small correlations suggesting that a large proportion of the pupil variance remains unexplained (Joshi et al., 2016; Megemont et al., 2022; Reimer et al., 2016).

The idea that the ongoing pupil signal fluctuations and the task-related pupil evoked response can be two largely independent signals, challenges previous claims. The prevailing model suggests that high baseline (tonic) activity in the LC is associated with weaker task-related (phasic) responses, and that pupil data parallels this observation (Aston-Jones and Cohen, 2005; Gilzenrat et al., 2010). However, the application of this model to human pupil data acquired during performance of typical psychophysical tasks has been challenged given the relatively small modulations in arousal that occur under these conditions (Joshi and Gold, 2020). Our analyses further suggest that the effect of pupil baseline on the pupil dilation responses is not behaviorally relevant and merely additive.

Previous fMRI studies suggest that task evoked responses are affected by ongoing signal fluctuations occurring in the same brain areas but that these ongoing fluctuations are shared across brain regions (Fox et al., 2006; Schölvinck et al., 2012). Similarly, single-cell electrophysiology and optical imaging with voltage sensors in anesthetized cats suggest that sensory responses in individual neurons are variable from trial-to-trial and that this variability reflects the sum of the sensory response with an ongoing signal shared across neurons (Arieli et al., 1996; Schölvinck et al., 2015). The CNV preparatory potential originates not from an unitary neural mechanism but from a set of brain areas, including the supplementary motor area (SMA), anterior cingulate cortex (ACC) and the thalamus (Nagai et al., 2004). Likely, these areas will display ongoing activity that will sum to the evoked responses. Nevertheless, the ongoing EEG fluctuations studied probably originate in a large set of brain areas that will include, but not be limited to, the brain areas involved in the preparatory mechanisms linked to the CNV potential.

The two markers of preparatory neural activity, the CNV and the pupil dilation response, were found to be correlated on a trial-by-trial basis. The correlation coefficients were stronger after adjusting for the slow ongoing fluctuations of each signal, further highlighting the importance of adjusting for these signal fluctuations. A correlation between the CNV and the pupil dilation response had already been suggested before on between-subjects correlation analysis (Hepsomali et al., 2019). We found that the correlation was stronger when comparing the CNV amplitude with the pupil dilation measured with a 500 ms delay in line with the idea that the pupil response is slow and lags the associated cortical events by hundreds of milliseconds (Reimer et al., 2016). A likely brain area linking the CNV with the pupil dilation response is the anterior cingulate cortex (ACC). Activity in the ACC is observed during preparatory periods associated with the CNV potential (Nagai et al., 2004) and it is associated with pupil dilation events (Joshi et al., 2016; Joshi and Gold, 2022). Nevertheless, it is intriguing to note that the correlation between the CNV and the pupil dilation response presented a clear posterior topography challenging the focus solely on the ACC. The fact that the CNV and pupil dilation independently explain some of the reaction time variance suggests the existence of brain systems that contribute towards the CNV amplitude independently of the brain systems associated with pupil dilation. Given that, without adjusting the CNV for the effect of the ongoing signal fluctuations, the correlation between the CNV and reaction time was observed mostly in left frontocentral channels, it is likely that this reflects the effect of a component of the CNV associated with motor preparatory activity independent of the orienting response. In turn, it is also intriguing that pupil dilation contributes towards reaction time independently of its effect on the CNV. Possible mechanisms are the priming of sensory cortices or the speeding up of the communication between sensory and motor cortices by the ascending brainstem neuromodulatory systems (Vazey et al., 2018).

Our analyses showed that ageing affects the dynamics of the ongoing EEG signal (reflecting cortical activity) and the ongoing pupillary signal (reflecting activity in brainstem neuromodulatory systems). In both signals, older people presented flatter spectra with reduced power in the slower aperiodic fluctuations. This flattening of brain signal spectra with ageing has been described before in EEG and ECoG data (Voytek et al., 2015; Waschke et al., 2017), is also evident in MEG data (Vlahou et al., 2014), and is here reported in the pupil signal suggesting an age-related decrease in the amplitude of the slow fluctuations in cortical and subcortical networks. In line with our findings, Vlahou et al., 2014 showed that, in MEG data, the power of slow fluctuations decreases linearly with ageing (Vlahou et al., 2014). Sequential activation and de-activation of resting-state networks is observed in the ongoing EEG data both during rest and task performance (Abreu et al., 2020; Liu et al., 2017). This ongoing dynamic possibly underlies the ongoing signal fluctuations we observe at the channel level EEG. The fact that EEG, MEG and pupil spectra decrease in amplitude with increasing frequency suggests that brain activation patterns that involve large activation volumes accumulate relatively slowly (the larger the brain volume recruited the slower the activity changes). Ageing is characterized by a significant decrease in the strength of long-range structural connections (Westlye et al., 2010) as well as a decrease in the functional connectivity (correlated BOLD activation patterns) within large-scale brain systems, like the default mode network (Sala-Llonch et al., 2015), consistent with the idea that simultaneous activation of distant brain areas supported by long-range structural connectivity is impaired in older individuals (Andrews-Hanna et al., 2007). Reduced amplitude of slow EEG fluctuations might therefore be related to impaired recruitment of large-scale brain systems.

We started this study with the assumption that if behavior is more variable in older people, then the neural responses that underlie that behavior should also show increased variability. We found that after controlling for differences in the ongoing brain signals, task-related evoked responses were equally variable in young and older adults. This leaves unanswered the origin of the increased behavioral variability observed in older people. It is likely that behavioral variability emerges from the summation of variability at different stages of task processing and thus it is not all captured by the preparatory responses studied here. Moreover, in our study, behavioral variability was only slightly increased in the older group and that might have hindered the detection of increased variability in the evoked responses. The age-related increase in reaction time variability depends on the task used to measure it (Dykiert et al., 2012). For example, the age difference in reaction time variability is smaller in simple RT tasks than in choice tasks and is smaller in tasks using longer preparatory intervals as the ones used in our task (Dykiert et al., 2012). Future studies addressing this question would benefit from using a behavioral task where age differences in behavioral variability are more pronounced.

The dataset used in this study has been made available in Open Neuro Repository under the https://doi.org/10.18112/openneuro.ds003690.v1.0.0. All analyses and figures presented in the manuscript resulted from analyses of this dataset. Matlab code used in the analyses is available on GitHub https://github.com/CIBIT-ICNAS/2021_Ribeiro_BrainVariability_Aging, (copy archived at swh:1:rev:e26e2e2ca77e95bdd1db48cb850273e6fd2c9940).

1. 1. Maria JR

   (2021) OpenNeuro Repository

   EEG, ECG and pupil data from young and older adults: rest and auditory cued reaction time tasks.

   https://doi.org/10.18112/openneuro.ds003690.v1.0.0

1. 1. Abreu R
   2. Simões M
   3. Castelo-Branco M

   (2020) Pushing the Limits of EEG: Estimation of Large-Scale Functional Brain Networks and Their Dynamics Validated by Simultaneous fMRI

   Frontiers in Neuroscience 14:323.

   https://doi.org/10.3389/fnins.2020.00323

   • PubMed
   • Google Scholar
2. 1. Andrews-Hanna JR
   2. Snyder AZ
   3. Vincent JL
   4. Lustig C
   5. Head D
   6. Raichle ME
   7. Buckner RL

   (2007) Disruption of large-scale brain systems in advanced aging

   Neuron 56:924–935.

   https://doi.org/10.1016/j.neuron.2007.10.038

   • PubMed
   • Google Scholar
3. 1. Arieli A
   2. Sterkin A
   3. Grinvald A
   4. Aertsen A

   (1996) Dynamics of ongoing activity: explanation of the large variability in evoked cortical responses

   Science (New York, N.Y.) 273:1868–1871.

   https://doi.org/10.1126/science.273.5283.1868

   • PubMed
   • Google Scholar
4. 1. Aston-Jones G
   2. Cohen JD

   (2005) An integrative theory of locus coeruleus-norepinephrine function: adaptive gain and optimal performance

   Annual Review of Neuroscience 28:403–450.

   https://doi.org/10.1146/annurev.neuro.28.061604.135709

   • PubMed
   • Google Scholar
5. 1. Becker R
   2. Reinacher M
   3. Freyer F
   4. Villringer A
   5. Ritter P

   (2011) How ongoing neuronal oscillations account for evoked fMRI variability

   The Journal of Neuroscience 31:11016–11027.

   https://doi.org/10.1523/JNEUROSCI.0210-11.2011

   • PubMed
   • Google Scholar
6. 1. Beckmann CF
   2. DeLuca M
   3. Devlin JT
   4. Smith SM

   (2005) Investigations into resting-state connectivity using independent component analysis

   Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 360:1001–1013.

   https://doi.org/10.1098/rstb.2005.1634

   • PubMed
   • Google Scholar
7. 1. Blair RC
   2. Karniski W

   (1993) An alternative method for significance testing of waveform difference potentials

   Psychophysiology 30:518–524.

   https://doi.org/10.1111/j.1469-8986.1993.tb02075.x

   • PubMed
   • Google Scholar
8. 1. Boehm U
   2. van Maanen L
   3. Forstmann B
   4. van Rijn H

   (2014) Trial-by-trial fluctuations in CNV amplitude reflect anticipatory adjustment of response caution

   NeuroImage 96:95–105.

   https://doi.org/10.1016/j.neuroimage.2014.03.063

   • PubMed
   • Google Scholar
9. 1. Boly M
   2. Balteau E
   3. Schnakers C
   4. Degueldre C
   5. Moonen G
   6. Luxen A
   7. Phillips C
   8. Peigneux P
   9. Maquet P
   10. Laureys S

   (2007) Baseline brain activity fluctuations predict somatosensory perception in humans

   PNAS 104:12187–12192.

   https://doi.org/10.1073/pnas.0611404104

   • PubMed
   • Google Scholar
10. 1. Boylan MA
    2. Foster CM
    3. Pongpipat EE
    4. Webb CE
    5. Rodrigue KM
    6. Kennedy KM

    (2021) Greater BOLD Variability is Associated With Poorer Cognitive Function in an Adult Lifespan Sample

    Cerebral Cortex (New York, N.Y 31:562–574.

    https://doi.org/10.1093/cercor/bhaa243

    • PubMed
    • Google Scholar
11. 1. Brainard DH

    (1997)

    The Psychophysics Toolbox

    Spatial Vision 10:433–436.

    • PubMed
    • Google Scholar
12. 1. Crameri F
    2. Shephard GE
    3. Heron PJ

    (2020) The misuse of colour in science communication

    Nature Communications 11:5444.

    https://doi.org/10.1038/s41467-020-19160-7

    • PubMed
    • Google Scholar
13. 1. de Gee JW
    2. Colizoli O
    3. Kloosterman NA
    4. Knapen T
    5. Nieuwenhuis S
    6. Donner TH

    (2017) Dynamic modulation of decision biases by brainstem arousal systems

    eLife 6:32.

    https://doi.org/10.7554/eLife.23232

    • PubMed
    • Google Scholar
14. 1. Delorme A
    2. Makeig S

    (2004) EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis

    Journal of Neuroscience Methods 134:9–21.

    https://doi.org/10.1016/j.jneumeth.2003.10.009

    • PubMed
    • Google Scholar
15. 1. Donoghue T
    2. Haller M
    3. Peterson EJ
    4. Varma P
    5. Sebastian P
    6. Gao R
    7. Noto T
    8. Lara AH
    9. Wallis JD
    10. Knight RT
    11. Shestyuk A
    12. Voytek B

    (2020) Parameterizing neural power spectra into periodic and aperiodic components

    Nature Neuroscience 23:1655–1665.

    https://doi.org/10.1038/s41593-020-00744-x

    • PubMed
    • Google Scholar
16. 1. Dykiert D
    2. Der G
    3. Starr JM
    4. Deary IJ

    (2012) Age differences in intra-individual variability in simple and choice reaction time: systematic review and meta-analysis

    PLOS ONE 7:e45759.

    https://doi.org/10.1371/journal.pone.0045759

    • PubMed
    • Google Scholar
17. 1. Fox MD
    2. Snyder AZ
    3. Zacks JM
    4. Raichle ME

    (2006) Coherent spontaneous activity accounts for trial-to-trial variability in human evoked brain responses

    Nature Neuroscience 9:23–25.

    https://doi.org/10.1038/nn1616

    • PubMed
    • Google Scholar
18. 1. Fox MD
    2. Snyder AZ
    3. Vincent JL
    4. Raichle ME

    (2007) Intrinsic fluctuations within cortical systems account for intertrial variability in human behavior

    Neuron 56:171–184.

    https://doi.org/10.1016/j.neuron.2007.08.023

    • PubMed
    • Google Scholar
19. 1. Garrett DD
    2. Nagel IE
    3. Preuschhof C
    4. Burzynska AZ
    5. Marchner J
    6. Wiegert S
    7. Jungehülsing GJ
    8. Nyberg L
    9. Villringer A
    10. Li SC
    11. Heekeren HR
    12. Bäckman L
    13. Lindenberger U

    (2015) Amphetamine modulates brain signal variability and working memory in younger and older adults

    PNAS 112:7593–7598.

    https://doi.org/10.1073/pnas.1504090112

    • PubMed
    • Google Scholar
20. 1. Gilzenrat MS
    2. Nieuwenhuis S
    3. Jepma M
    4. Cohen JD

    (2010) Pupil diameter tracks changes in control state predicted by the adaptive gain theory of locus coeruleus function

    Cognitive, Affective & Behavioral Neuroscience 10:252–269.

    https://doi.org/10.3758/CABN.10.2.252

    • PubMed
    • Google Scholar
21. 1. Grady CL
    2. Garrett DD

    (2014) Understanding variability in the BOLD signal and why it matters for aging

    Brain Imaging and Behavior 8:274–283.

    https://doi.org/10.1007/s11682-013-9253-0

    • PubMed
    • Google Scholar
22. 1. Grady CL
    2. Garrett DD

    (2018) Brain signal variability is modulated as a function of internal and external demand in younger and older adults

    NeuroImage 169:510–523.

    https://doi.org/10.1016/j.neuroimage.2017.12.031

    • PubMed
    • Google Scholar
23. 1. Groppe DM
    2. Urbach TP
    3. Kutas M

    (2011) Mass univariate analysis of event-related brain potentials/fields I: A critical tutorial review

    Psychophysiology 48:1711–1725.

    https://doi.org/10.1111/j.1469-8986.2011.01273.x

    • PubMed
    • Google Scholar
24. Website

    1. Groppe DM

    (2021) MATLAB Central File Exchange

    Accessed April 20, 2021.

    https://www.mathworks.com/matlabcentral/fileexchange/29782-mult_comp_perm_t1-data-n_perm-tail-alpha_level-mu-reports-seed_state
25. 1. Hepsomali P
    2. Hadwin JA
    3. Liversedge SP
    4. Degno F
    5. Garner M

    (2019) The impact of cognitive load on processing efficiency and performance effectiveness in anxiety: evidence from event-related potentials and pupillary responses

    Experimental Brain Research 237:897–909.

    https://doi.org/10.1007/s00221-018-05466-y

    • Google Scholar
26. 1. Hong L
    2. Walz JM
    3. Sajda P

    (2014) Your eyes give you away: prestimulus changes in pupil diameter correlate with poststimulus task-related EEG dynamics

    PLOS ONE 9:e91321.

    https://doi.org/10.1371/journal.pone.0091321

    • PubMed
    • Google Scholar
27. 1. Jennings JR
    2. van der Molen MW
    3. Steinhauer SR

    (1998) Preparing the heart, eye, and brain: foreperiod length effects in a nonaging paradigm

    Psychophysiology 35:90–98.

    https://doi.org/10.1111/1469-8986.3510090

    • PubMed
    • Google Scholar
28. 1. Joshi S
    2. Li Y
    3. Kalwani RM
    4. Gold JI

    (2016) Relationships between Pupil Diameter and Neuronal Activity in the Locus Coeruleus, Colliculi, and Cingulate Cortex

    Neuron 89:221–234.

    https://doi.org/10.1016/j.neuron.2015.11.028

    • PubMed
    • Google Scholar
29. 1. Joshi S
    2. Gold JI

    (2020) Pupil Size as a Window on Neural Substrates of Cognition

    Trends in Cognitive Sciences 24:466–480.

    https://doi.org/10.1016/j.tics.2020.03.005

    • PubMed
    • Google Scholar
30. 1. Joshi S
    2. Gold JI

    (2022) Context-dependent relationships between locus coeruleus firing patterns and coordinated neural activity in the anterior cingulate cortex

    eLife 11:90.

    https://doi.org/10.7554/eLife.63490

    • PubMed
    • Google Scholar
31. 1. Kosciessa JQ
    2. Kloosterman NA
    3. Garrett DD

    (2020) Standard multiscale entropy reflects neural dynamics at mismatched temporal scales: What’s signal irregularity got to do with it?

    PLOS Computational Biology 16:e1007885.

    https://doi.org/10.1371/journal.pcbi.1007885

    • PubMed
    • Google Scholar
32. 1. Kumral D
    2. Şansal F
    3. Cesnaite E
    4. Mahjoory K
    5. Al E
    6. Gaebler M
    7. Nikulin VV
    8. Villringer A

    (2020) BOLD and EEG signal variability at rest differently relate to aging in the human brain

    NeuroImage 207:116373.

    https://doi.org/10.1016/j.neuroimage.2019.116373

    • PubMed
    • Google Scholar
33. 1. Liu Q
    2. Farahibozorg S
    3. Porcaro C
    4. Wenderoth N
    5. Mantini D

    (2017) Detecting large-scale networks in the human brain using high-density electroencephalography

    Human Brain Mapping 38:4631–4643.

    https://doi.org/10.1002/hbm.23688

    • PubMed
    • Google Scholar
34. 1. Lou B
    2. Li Y
    3. Philiastides MG
    4. Sajda P

    (2014) Prestimulus alpha power predicts fidelity of sensory encoding in perceptual decision making

    NeuroImage 87:242–251.

    https://doi.org/10.1016/j.neuroimage.2013.10.041

    • PubMed
    • Google Scholar
35. 1. MacDonald SWS
    2. Karlsson S
    3. Rieckmann A
    4. Nyberg L
    5. Bäckman L

    (2012) Aging-related increases in behavioral variability: relations to losses of dopamine D1 receptors

    The Journal of Neuroscience 32:8186–8191.

    https://doi.org/10.1523/JNEUROSCI.5474-11.2012

    • PubMed
    • Google Scholar
36. 1. Mayhew SD
    2. Ostwald D
    3. Porcaro C
    4. Bagshaw AP

    (2013) Spontaneous EEG alpha oscillation interacts with positive and negative BOLD responses in the visual-auditory cortices and default-mode network

    NeuroImage 76:362–372.

    https://doi.org/10.1016/j.neuroimage.2013.02.070

    • PubMed
    • Google Scholar
37. 1. McIntosh AR
    2. Vakorin V
    3. Kovacevic N
    4. Wang H
    5. Diaconescu A
    6. Protzner AB

    (2014) Spatiotemporal dependency of age-related changes in brain signal variability

    Cerebral Cortex (New York, N.Y 24:1806–1817.

    https://doi.org/10.1093/cercor/bht030

    • PubMed
    • Google Scholar
38. 1. Megemont M
    2. McBurney-Lin J
    3. Yang H

    (2022) Pupil diameter is not an accurate real-time readout of locus coeruleus activity

    eLife 11:510.

    https://doi.org/10.7554/eLife.70510

    • PubMed
    • Google Scholar
39. 1. Monto S
    2. Palva S
    3. Voipio J
    4. Palva JM

    (2008) Very slow EEG fluctuations predict the dynamics of stimulus detection and oscillation amplitudes in humans

    The Journal of Neuroscience 28:8268–8272.

    https://doi.org/10.1523/JNEUROSCI.1910-08.2008

    • PubMed
    • Google Scholar
40. 1. Murphy PR
    2. O’Connell RG
    3. O’Sullivan M
    4. Robertson IH
    5. Balsters JH

    (2014) Pupil diameter covaries with BOLD activity in human locus coeruleus

    Human Brain Mapping 35:4140–4154.

    https://doi.org/10.1002/hbm.22466

    • PubMed
    • Google Scholar
41. 1. Nagai Y
    2. Critchley HD
    3. Featherstone E
    4. Fenwick PBC
    5. Trimble MR
    6. Dolan RJ

    (2004) Brain activity relating to the contingent negative variation: an fMRI investigation

    NeuroImage 21:1232–1241.

    https://doi.org/10.1016/j.neuroimage.2003.10.036

    • PubMed
    • Google Scholar
42. 1. Nomi JS
    2. Bolt TS
    3. Ezie CEC
    4. Uddin LQ
    5. Heller AS

    (2017) Moment-to-Moment BOLD Signal Variability Reflects Regional Changes in Neural Flexibility across the Lifespan

    The Journal of Neuroscience 37:5539–5548.

    https://doi.org/10.1523/JNEUROSCI.3408-16.2017

    • PubMed
    • Google Scholar
43. 1. Nurhab MI
    2. Nurhab B
    3. Purwaningsih T
    4. Teng MF

    (2017) Circular(2)-linear regression analysis with iteration order manipulation

    International Journal of Advances in Intelligent Informatics 3:107.

    https://doi.org/10.26555/ijain.v3i2.90

    • Google Scholar
44. 1. Pernet CR
    2. Wilcox R
    3. Rousselet GA

    (2012) Robust correlation analyses: false positive and power validation using a new open source matlab toolbox

    Frontiers in Psychology 3:1–18.

    https://doi.org/10.3389/fpsyg.2012.00606

    • PubMed
    • Google Scholar
45. 1. Podvalny E
    2. King LE
    3. He BJ

    (2021) Spectral signature and behavioral consequence of spontaneous shifts of pupil-linked arousal in human

    eLife 10:65.

    https://doi.org/10.7554/eLife.68265

    • PubMed
    • Google Scholar
46. 1. Reimer J
    2. McGinley MJ
    3. Liu Y
    4. Rodenkirch C
    5. Wang Q
    6. McCormick DA
    7. Tolias AS

    (2016) Pupil fluctuations track rapid changes in adrenergic and cholinergic activity in cortex

    Nature Communications 7:1–7.

    https://doi.org/10.1038/ncomms13289

    • PubMed
    • Google Scholar
47. 1. Ribeiro MJ
    2. Paiva JS
    3. Castelo-Branco M

    (2016) Spontaneous Fluctuations in Sensory Processing Predict Within-Subject Reaction Time Variability

    Frontiers in Human Neuroscience 10:1–15.

    https://doi.org/10.3389/fnhum.2016.00200

    • PubMed
    • Google Scholar
48. 1. Ribeiro MJ
    2. Castelo-Branco M

    (2019a) Age-related differences in event-related potentials and pupillary responses in cued reaction time tasks

    Neurobiology of Aging 73:177–189.

    https://doi.org/10.1016/j.neurobiolaging.2018.09.028

    • PubMed
    • Google Scholar
49. 1. Ribeiro MJ
    2. Castelo-Branco M

    (2019b) Neural correlates of anticipatory cardiac deceleration and its association with the speed of perceptual decision-making, in young and older adults

    NeuroImage 199:521–533.

    https://doi.org/10.1016/j.neuroimage.2019.06.004

    • PubMed
    • Google Scholar
50. Website

    1. Ribeiro MJ
    2. Castelo-Branco M

    (2021) EEG, ECG and pupil data from young and older adults: rest and auditory cued reaction time tasks

    OpenNeuro. Accessed June 10, 2021.

    https://doi.org/10.18112/openneuro.ds003690.v1.0.0
51. 1. Rousseeuw PJ

    (1984) Least Median of Squares Regression

    Journal of the American Statistical Association 79:871–880.

    https://doi.org/10.1080/01621459.1984.10477105

    • Google Scholar
52. 1. Sala-Llonch R
    2. Bartrés-Faz D
    3. Junqué C

    (2015) Reorganization of brain networks in aging: a review of functional connectivity studies

    Frontiers in Psychology 6:1–11.

    https://doi.org/10.3389/fpsyg.2015.00663

    • PubMed
    • Google Scholar
53. 1. Samanez-Larkin GR
    2. Kuhnen CM
    3. Yoo DJ
    4. Knutson B

    (2010) Variability in nucleus accumbens activity mediates age-related suboptimal financial risk taking

    The Journal of Neuroscience 30:1426–1434.

    https://doi.org/10.1523/JNEUROSCI.4902-09.2010

    • PubMed
    • Google Scholar
54. 1. Sander MC
    2. Werkle-Bergner M
    3. Lindenberger U

    (2012) Amplitude modulations and inter-trial phase stability of alpha-oscillations differentially reflect working memory constraints across the lifespan

    NeuroImage 59:646–654.

    https://doi.org/10.1016/j.neuroimage.2011.06.092

    • PubMed
    • Google Scholar
55. 1. Schölvinck ML
    2. Friston KJ
    3. Rees G

    (2012) The influence of spontaneous activity on stimulus processing in primary visual cortex

    NeuroImage 59:2700–2708.

    https://doi.org/10.1016/j.neuroimage.2011.10.066

    • PubMed
    • Google Scholar
56. 1. Schölvinck ML
    2. Saleem AB
    3. Benucci A
    4. Harris KD
    5. Carandini M

    (2015) Cortical state determines global variability and correlations in visual cortex

    The Journal of Neuroscience 35:170–178.

    https://doi.org/10.1523/JNEUROSCI.4994-13.2015

    • PubMed
    • Google Scholar
57. 1. Shimaoka D
    2. Steinmetz NA
    3. Harris KD
    4. Carandini M

    (2019) The impact of bilateral ongoing activity on evoked responses in mouse cortex

    eLife 8:33.

    https://doi.org/10.7554/eLife.43533

    • PubMed
    • Google Scholar
58. 1. Siegle GJ
    2. Steinhauer SR
    3. Stenger VA
    4. Konecky R
    5. Carter CS

    (2003) Use of concurrent pupil dilation assessment to inform interpretation and analysis of fMRI data

    NeuroImage 20:114–124.

    https://doi.org/10.1016/s1053-8119(03)00298-2

    • PubMed
    • Google Scholar
59. 1. Smith SM
    2. Fox PT
    3. Miller KL
    4. Glahn DC
    5. Fox PM
    6. Mackay CE
    7. Filippini N
    8. Watkins KE
    9. Toro R
    10. Laird AR
    11. Beckmann CF

    (2009) Correspondence of the brain’s functional architecture during activation and rest

    PNAS 106:13040–13045.

    https://doi.org/10.1073/pnas.0905267106

    • PubMed
    • Google Scholar
60. 1. Thut G
    2. Nietzel A
    3. Brandt SA
    4. Pascual-Leone A

    (2006) Alpha-band electroencephalographic activity over occipital cortex indexes visuospatial attention bias and predicts visual target detection

    The Journal of Neuroscience 26:9494–9502.

    https://doi.org/10.1523/JNEUROSCI.0875-06.2006

    • PubMed
    • Google Scholar
61. 1. Tran TT
    2. Hoffner NC
    3. LaHue SC
    4. Tseng L
    5. Voytek B

    (2016) Alpha phase dynamics predict age-related visual working memory decline

    NeuroImage 143:196–203.

    https://doi.org/10.1016/j.neuroimage.2016.08.052

    • PubMed
    • Google Scholar
62. 1. Tran TT
    2. Rolle CE
    3. Gazzaley A
    4. Voytek B

    (2020) Linked Sources of Neural Noise Contribute to Age-related Cognitive Decline

    Journal of Cognitive Neuroscience 32:1813–1822.

    https://doi.org/10.1162/jocn_a_01584

    • PubMed
    • Google Scholar
63. 1. Tsvetanov KA
    2. Henson RNA
    3. Jones PS
    4. Mutsaerts H
    5. Fuhrmann D
    6. Tyler LK
    7. Cam-CAN
    8. Rowe JB

    (2021) The effects of age on resting-state BOLD signal variability is explained by cardiovascular and cerebrovascular factors

    Psychophysiology 58:1–20.

    https://doi.org/10.1111/psyp.13714

    • PubMed
    • Google Scholar
64. 1. van Boxtel GJM
    2. Böcker KBE

    (2004) Cortical Measures of Anticipation

    Journal of Psychophysiology 18:61–76.

    https://doi.org/10.1027/0269-8803.18.23.61

    • Google Scholar
65. 1. van den Brink RL
    2. Murphy PR
    3. Nieuwenhuis S

    (2016) Pupil Diameter Tracks Lapses of Attention

    PLOS ONE 11:16.

    https://doi.org/10.1371/journal.pone.0165274

    • PubMed
    • Google Scholar
66. 1. van Diepen RM
    2. Cohen MX
    3. Denys D
    4. Mazaheri A

    (2015) Attention and Temporal Expectations Modulate Power, Not Phase, of Ongoing Alpha Oscillations

    Journal of Cognitive Neuroscience 27:1573–1586.

    https://doi.org/10.1162/jocn_a_00803

    • PubMed
    • Google Scholar
67. 1. van Dijk H
    2. Schoffelen J-M
    3. Oostenveld R
    4. Jensen O

    (2008) Prestimulus Oscillatory Activity in the Alpha Band Predicts Visual Discrimination Ability

    Journal of Neuroscience 28:1816–1823.

    https://doi.org/10.1523/JNEUROSCI.1853-07.2008

    • PubMed
    • Google Scholar
68. 1. Vazey EM
    2. Moorman DE
    3. Aston-Jones G

    (2018) Phasic locus coeruleus activity regulates cortical encoding of salience information

    PNAS 115:40.

    https://doi.org/10.1073/pnas.1803716115

    • PubMed
    • Google Scholar
69. 1. Vlahou EL
    2. Thurm F
    3. Kolassa IT
    4. Schlee W

    (2014) Resting-state slow wave power, healthy aging and cognitive performance

    Scientific Reports 4:5101.

    https://doi.org/10.1038/srep05101

    • PubMed
    • Google Scholar
70. 1. Voytek B
    2. Kramer MA
    3. Case J
    4. Lepage KQ
    5. Tempesta ZR
    6. Knight RT
    7. Gazzaley A

    (2015) Age-Related Changes in 1/f Neural Electrophysiological Noise

    Journal of Neuroscience 35:13257–13265.

    https://doi.org/10.1523/JNEUROSCI.2332-14.2015

    • PubMed
    • Google Scholar
71. 1. Wang CA
    2. Boehnke SE
    3. Itti L
    4. Munoz DP

    (2014) Transient pupil response is modulated by contrast-based saliency

    The Journal of Neuroscience 34:408–417.

    https://doi.org/10.1523/JNEUROSCI.3550-13.2014

    • PubMed
    • Google Scholar
72. 1. Waschke L
    2. Wöstmann M
    3. Obleser J

    (2017) States and traits of neural irregularity in the age-varying human brain

    Scientific Reports 7:17381.

    https://doi.org/10.1038/s41598-017-17766-4

    • PubMed
    • Google Scholar
73. 1. Westlye LT
    2. Walhovd KB
    3. Dale AM
    4. Bjørnerud A
    5. Due-Tønnessen P
    6. Engvig A
    7. Grydeland H
    8. Tamnes CK
    9. Ostby Y
    10. Fjell AM

    (2010) Life-span changes of the human brain white matter: diffusion tensor imaging (DTI) and volumetry

    Cerebral Cortex (New York, N.Y 20:2055–2068.

    https://doi.org/10.1093/cercor/bhp280

    • PubMed
    • Google Scholar
74. 1. Wiegand I
    2. Sander MC

    (2019) Cue-related processing accounts for age differences in phasic alerting

    Neurobiology of Aging 79:93–100.

    https://doi.org/10.1016/j.neurobiolaging.2019.03.017

    • PubMed
    • Google Scholar
75. Book

    1. Wilcox RR

    (2012)

    Introduction to Robust Estimation and Hypothesis Testing (Edn (3rd ed)

    Academic Press.

    • Google Scholar
76. 1. Yang H
    2. Bari BA
    3. Cohen JY
    4. O’Connor DH

    (2021) Locus coeruleus spiking differently correlates with S1 cortex activity and pupil diameter in a tactile detection task

    eLife 10:327.

    https://doi.org/10.7554/eLife.64327

    • Google Scholar

Author details

1. Maria Ribeiro

   1. CIBIT-ICNAS, University of Coimbra, Coimbra, Portugal
   2. Faculty of Medicine, University of Coimbra, Coimbra, Portugal

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Visualization, Writing – original draft, Writing – review and editing

   For correspondence

   mjribeiro@fmed.uc.pt

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-6422-3279

2. Miguel Castelo-Branco

   1. CIBIT-ICNAS, University of Coimbra, Coimbra, Portugal
   2. Faculty of Medicine, University of Coimbra, Coimbra, Portugal

   Contribution

   Funding acquisition, Resources, Supervision, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4364-6373

• 1,023

  views

• 136

  downloads

• 4

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.