Multiple decades of neuroscience and psychology research uncovered that arousal contributes to enhanced memory performance and that norepinephrine (NE) transmission plays a pivotal role in this phenomenon (van Stegeren, 2008; Roozendaal and McGaugh, 2011). The major source of NE to the brain is the locus coeruleus (LC), a small structure in the brainstem. The NE-LC system has widespread efferents to almost the entire brain and can modulate cognition, behavior and autonomic tone via its effects on adrenoreceptors in target neurons (Sara, 2009; Sara and Bouret, 2012; Samuels and Szabadi, 2008a).

Research into NE-modulation of emotional memory enhancement has focused largely on the amygdala (Hermans et al., 2014; Phelps and LeDoux, 2005; McIntyre et al., 2003), an important target region of the LC and a critical nexus in memory for arousing experiences. Rodent work consistently demonstrated that blocking β-adrenergic receptors in the basolateral amygdala (BLA) immediately after learning in arousing conditions impairs memory, while infusion of NE or β-adrenoreceptor agonists enhances consolidation and rescinds the impairment, evaluated one to two days later (Roozendaal et al., 2008; Power et al., 2002). Correspondingly, stimulation of the LC during the learning phase modulates activity in the BLA and synaptic plasticity in the CA1, CA3 and dentate gyrus (Lemon et al., 2009; Wagatsuma et al., 2018). While these studies highlight the importance of adrenergic signaling for plasticity and hippocampal-dependent learning, blocking β-adrenoreceptors prior to learning had no effect on subsequent retrieval 24 hr later (Wagatsuma et al., 2018; Khakpour-Taleghani et al., 2008). It has been suggested that in the context of strong LC stimulation tonic activity can be shifted into phasic activity and other neuromodulators may then elicit similar but possibly weaker plastic processes (Roozendaal and Hermans, 2017; Vazey et al., 2018; Hansen and Manahan-Vaughan, 2015).

In humans, administration of propranolol 90 min before encoding attenuated expected amygdala responses during encoding of emotional stimuli, and even though the propranolol was eliminated at retrieval, hippocampal responses during retrieval of emotional stimuli were absent (Strange and Dolan, 2004). In addition, propranolol impaired memory accuracy for emotional events and decreased saliva measures of NE (Hermans et al., 2014; McIntyre et al., 2003; Strange and Dolan, 2004; Cahill et al., 1994). These findings suggest, consistent with animal studies, that emotional memory is dependent upon β-adrenergic-related amygdala-hippocampus interactions.

These network interactions can vary over time or be task-dependent. For example, the LC is involved in consolidation, but within a critical time period after learning, the time required for the hippocampus to facilitate the reorganization and stabilization of storing information in the neocortex (Khakpour-Taleghani et al., 2008). In humans, increases in NE-tone during encoding, but not during retrieval, modulated memory performance (Rimmele et al., 2016). This dynamic behavior stems from the idea that release of NE due to LC-activation modulates, gates and tunes neural activity in a way that optimizes the signal-to-noise ratio (Sara, 2009; Sara and Bouret, 2012). Thus, NE selectively reinforces brain activity and memory representations only if arousal-related LC activation occurs at the right time and magnitude (Mather et al., 2016). When increases in NE are coupled to arousal, they can lead to hippocampal synaptic plasticity (Lemon et al., 2009; Hansen and Manahan-Vaughan, 2015), and facilitate the dynamic reorganization of neural networks (Bouret and Sara, 2005; Zerbi et al., 2019). So far, arousal-induced dynamic changes in LC activity and connectivity with its key target regions during task performance has not yet been examined in animals (Zerbi et al., 2019).

Motivated by the previous findings, we sought to examine the contribution of the LC directly in vivo in humans to arousal-related memory enhancement and hypothesized that interactions between the LC and MTL regions vary across memory stages, dependent on the level of arousal.

Examining the LC in vivo functionally and obtaining a robust signal in humans is exceedingly difficult given its small size, proneness to noise (pulsatility of surrounding vessels and motion) and its location adjacent to the fourth ventricle. Our recent efforts in ultra-high-field imaging, providing better spatial resolution and signal-to-noise ratio, open up exciting possibilities. Here, we combined 7T-fMRI scanning during an established emotional memory paradigm, our novel 7T-MRI sequence which allows to image the LC at high resolution (Priovoulos et al., 2018), with objective measures of arousal, viz. online physiological measures and serial saliva measures, a proxy for NE, in young adults.

Twenty-seven young right-handed adults (13 female, 20–30 years, mean ± SD = 22.95±1.96) underwent 7T MR imaging which included structural scans (T1-weighted scan, dedicated LC imaging) and ultra-high-resolution blood oxygenation level-dependent (BOLD) signal fMRI. The functional imaging part consisted of a baseline resting-state scan and an emotional memory task, in which participants were required to memorize face-name associations (Figure 1). Faces had an emotional (negative) or neutral expression as determined by independent rating. After the encoding phase, a second resting-state scan (termed ‘consolidation’) was collected. This was followed with a recollection fMRI-scan, during which participants had to indicate whether they recognized the face and, upon endorsement, had to select the name of the face out of three options (all options were seen previously to stimulate recollection instead of recognition processes). During the functional imaging, we collected breathing (respiratory bellows), cardiac information (pulse oximeter) and serial saliva samples for alpha-amylase (sAA) measures, a proxy for catecholaminergic activity. Three fMRI-datasets were excluded due to technical reasons, one due to spiking because of coil failure one due to a reconstruction error and one due to temporary failure of the response-recording device. Supplementary file 1 summarizes the neuropsychological test performance of the participants.

Figure 1 with 2 supplements see all

Download asset Open asset

Emotional memory performance is linked to sAA and heart-rate variability changes

We first evaluated the effect of emotional valence on memory performance, as it is well established that emotional information is better remembered than neutral information. All participants performed the task above chance level in recollection (33.3%), with a median number of hits (correctly recollected) of 48.33% (IQR = 40.78, 54.89) and median number of miss rate (incorrectly recollected) of 51.67% (IQR = 45.11, 59.22). There was no significant difference (Wilcoxon paired test: Z = −1.17, p-value=0.244) between the recognition hit rate for emotional (Median = 53.33% (IQR = 45.56, 61.11)) compared to neutral faces (Median = 53.33% (IQR = 44.44, 67.79, Figure 2A). However, more false alarms were detected (Wilcoxon paired test: Z = 3.81, p-value<o0.001) for emotional (Median = 22.73% (IQR = 18.19, 34.09) compared to neutral faces (Median = 13.63% (IQR = 9.09, 22.78); Figure 2B). We observed no recollection difference (Wilcox test: Z = 0.06, p-value=0.95, Figure 2D) between emotional (median = 0.5 (IQR = 0.43, 0.54)) and neutral valence trials (median = 0.45 (IQR = 0.38,0.59), Table 1). However, given that performance in the recollection phase may be affected by a response bias in the recognition phase (Figure 2E–F), we calculated the false alarm rate and the response bias using the likelihood ratio beta from signal detection theory. A paired t-test revealed a bias in misclassifying emotional faces during recognition (t(27)=5.34, p-value<0.001, Figure 2B) and a higher response bias for the emotional recognition condition (t(27)=2.49, p-value=0.019, Figure 2C), towards responding that they had seen the face before. We decided to focus on recollection because of this response bias, our focus on the medial temporal lobe regions and the notion that recollection is more likely to elicit hippocampal activity than recognition (Rugg and Vilberg, 2013).

Figure 2 with 1 supplement see all

Download asset Open asset

Figure 2—source data 1

Behavioral performance on the face-name association task.

Individual hit rates, false alarms, bias scores for the recognition, recollection scores (raw and bias-adjusted) are provided for the neutral and emotional valence condition.

https://cdn.elifesciences.org/articles/52059/elife-52059-fig2-data1-v2.rdata

Download elife-52059-fig2-data1-v2.rdata

Table 1

	Emotional (Median (IQR))	Neutral (Median (IQR))	Z	p-value
Recognition Hit Rate (Old faces)	53.33 (45.56, 61.11)	53.33 (44.44, 67.78)	−1.17	0.244
Recognition Miss Rate (Old faces)	46.67 (38.89, 54.44)	46.67 (32.22, 55.56)	1.2	0.229
Recognition Correct Rejection Rate (New faces)	77.27 (65.91, 81.82)	86.36 (77.27, 90.91)	−3.77	<0.001
Recognition False Alarm Rate (New faces)	22.73 (18.18, 34.09)	13.64 (9.09, 22.73)	3.81	<0.001
Recognition Hit Rate - Recognition False Alarm Rate	23.94 (17.02, 32.83)	37.68 (28.23, 50.81)	−3.16	<0.001
Recognition Response Bias	0.90 (0.81, 0.92)	0.84 (0.74, 0.9)	2.29	0.02
Recollection Hit Rate Raw	24.44 (18.89, 31.11)	26.67 (18.89, 33.33)	0.35	0.726
Recollection Hit Rate (correct recognition)	50 (43.43, 54.01)	44.83 (37.97, 58.77)	0.06	0.953
Recollection Hit Rate (Old faces) - Recollection False Alarm Rate (Old faces)	0.00 (−13.14–8.01)	−10.34 (−24.07–17.54)	0.06	0.953
Bias-corrected Recollection Hit Rate	38.68 (36.68, 39.46)	29.97 (28.01, 33.39)	4.41	<0.001

Repeating the analysis after correcting for the bias (partial residuals), showed that participants recollected on average more names with an emotional (median = 38.68, IQR = 36.68, 39.46) relative to those with a neutral valence (median = 29.97, IQR = 28.01, 33.39; paired Wilcoxon test: Z = 4.41, p-value<0.001, Figure 2G), in accordance with an expected memory advantage for emotional trials. No reaction time difference between emotional (median = 1.90 (IQR = 1.68,2.10)) and neutral conditions (median = 1.88 (IQR = 1.71, 2.02)) was detected in recollection (Z = 0.09, p-value=0.93) or recognition (Z = −0.43, p-value=0.68; emotional (median = 1.50 (IQR = 1.32, 1.65)); neutral (median = 1.52 (IQR = 1.32, 1.61))). In the remaining analyses, we used this bias-corrected emotional recollection score (Figure 2—figure supplement 1; Table 1 for an overview of all scores). Males and females did no differ in the corrected recollected rate (β = 1.49, z = 1.18, p=0.24), nor did we observe a possible modulation by valence (interaction sex by valence: β = 0.20, z = 0.91, p=0.91).

For task clarification, see Figure 1. The ‘Recollection Hit Rate Raw’ expresses the number of correct recollection of old names divided by the total number of old faces. The ‘Recollection Hit Rate (correct recognition)’ expresses the recollection success normalized for recognition, that is the number of correct recollection of old names divided by the number of correct recognition of old faces (%Hits). The %False Alarm is 1- %Hits (or Recollection Hit Rate (correct recognition)”). The ‘Bias-corrected Recollection Hit Rate’ is the bias-corrected recognition rate, as described in Figure 2—figure supplement 1. Note that the recollection hit rate for new faces is by definition 0, since they have not been shown in association with a name during encoding.

While there is general agreement in the valence of these images, individual differences in the elicited arousal are common. Therefore, we collected autonomic tone measures. We recorded heart rate variability (HRV), a proxy of heart rate variability and an index of parasympathetic modulation, particularly in the high (0.15–0.4 Hz) and low-frequency band (0.03–0.15 Hz) (Hedman et al., 1995). Root-mean square differences of successive peak-to-peak intervals (rMSSD) were calculated over each time period. RMSSD correlates tightly to parasympathetic activity (Kleiger et al., 2005) and is relatively free of respiratory influences. In addition, we measured sAA, a sensitive marker for the sympathetic nervous system, in particular NE.

To facilitate analyzing rMSSD and sAA at similar time points along the paradigm, we calculated the difference in sAA (ΔsAA) levels before and after each of the stages of the fMRI task (resulting in four values per participants, except for missing values due to insufficient saliva; Supplementary file 2). Linear mixed effects models with ΔsAA across the task stages revealed an increase in ΔsAA during consolidation relative to encoding (β = −65.95, t(47.92) = −2.91, p=0.033, Figure 3A). No differences in ΔsAA were found when comparing the other task stages (Supplementary file 3A). These findings did not change when covarying for the baseline sAA measurement outside the scanner or when covarying for sex. To further probe the increase in ΔsAA during consolidation, correlation analyses showed a positive, at trend level, association with emotional memory performance (r = 0.452, N = 18, p-value=0.059, Figure 3B).

Figure 3

Download asset Open asset

Figure 3—source data 1

Autonomic tone and memory performance across the entire paradigm.

Individual ΔsAA, rMSSD values per task stage. The emotional memory scores provided for panel B show the proportion of adjusted emotional scores versus neutral scores as indicated in the methods.

https://cdn.elifesciences.org/articles/52059/elife-52059-fig3-data1-v2.rdata

Download elife-52059-fig3-data1-v2.rdata

For the rMSSD, the linear mixed effects models showed no significant change in rMSSD across the task stages (Figure 3C, Supplementary file 3B). Given that both measures relate to the autonomic system, we then correlated ΔsAA to RMSSD across the stages using bootstrapped (n = 5000) repeated measures correlations and observed a negative relationship (R_rm(df)=−0.387,(35), p=0.018, CI (95%)=[−0.638,–0.061], Figure 3D). These observations suggest that perturbations of arousal with a NE-dominant response, reflected in increases in sympathetic and decreases in parasympathetic tone, can enhance memory performance in healthy adults.

LC activity couples with autonomic tone across the task stages

The LC is the centerpiece of the brain’s sympathetic noradrenergic system, and is known to control autonomic function. Therefore, we examined the relationship between LC fMRI and HRV timeseries across the task stages.

Using brainstem-specific group-level independent-component-analyses (ICA) on the baseline resting-state fMRI in combination with a custom anatomical LC template from our MT-TFL sequence, we identified a bilateral LC component (Figure 4A and B). To assess the specificity of our findings to the LC, we also investigated a control component in the pons (reference ROI). In addition, sensitivity analyses were done by evaluating the outcomes across four different preprocessing pipelines, each time adding in more physiological noise correction (Figure 4—figure supplement 1).

Figure 4 with 2 supplements see all

Download asset Open asset

Figure 4—source data 1

Locus coeruleus BOLD and autonomic tone measures across the entire paradigm.

Individual ΔsAA, rMSSD, LC and reference BOLD values per task stage. Individual coherence values between.

https://cdn.elifesciences.org/articles/52059/elife-52059-fig4-data1-v2.rdata

Download elife-52059-fig4-data1-v2.rdata

Linear mixed effects models showed no differences across task stages for LC BOLD variability (Supplementary file 3C). Consistent with our previous analyses, we applied repeated measures correlation and observed a negative correlation between LC variance and rMSSD across task stages (R_rm(df)=−0.326,(63), p-value=0.008, CI (95%)=[−0.531,–0.084], Figure 4C). We found no correlation between rMSSD and variance in a control region, a component close to the 4^th ventricle that explained a similar amount of variance in the ICA (R_rm(df)=−0.201(63), p=0.11, CI (95%)=[−0.494, 0.427], Figure 4D). We also observed no correlation between LC variance and ΔsAA (R_rm(df)=−0.006,(38), p-value=0.971, CI (95%)=[−0.325, 0.314], Figure 4E) or between the reference component variance and ΔsAA (R_rm(df)=−0.012,(38), p-value=0.943, CI (95%)=[−0.309, 0.330], Figure 4F). These findings were reproduced across preprocessing pipelines indicating the robustness of our results (Figure 4—figure supplement 2).

In the previous analyses, we demonstrated a negative correlation between LC variability and rMSSD or ΔsAA across the entire paradigm. These results echo previous work demonstrating that blocking β-adrenoreceptor increases HRV (Bittiner and Smith, 1986). However, rMSSD reflects beat-to-beat variance in HR in the time-domain and is relatively non-specific to the frequency-domain. Autonomic fluctuations due to emotionally salient stimuli can alter the power of frequencies in the HRV and this can occur in concert with changes in LC BOLD responses (Bittiner and Smith, 1986; Akselrod et al., 1981; Lane et al., 2009). Therefore, we applied spectral analyses to assess the nature of autonomic modulations on LC activity at each task stage by calculating the coherence magnitude squared between the timeseries of the HRV and LC for each task stage (see example in Figure 5A). By using a within-subject design with a baseline measure we can eliminate interindividual differences in blood pressure, respiration and, by combining HRV fluctuations with other measures related to the autonomic system, we are well-positioned to differentiate autonomic contributions to LC activity due to emotion or arousal across the different memory stages. Here, we will use the FIXed+explicit Phys pipeline to remove respiratory confounding.

Figure 5 with 3 supplements see all

Download asset Open asset

Figure 5—source data 1

Coherence between locus coeruleus BOLD and HRV.

Individual coherence values between locus coeruleus or reference BOLD and HRV per frequency are provided for each task stage in the paradigm. For panel E and F, the individual ΔsAA and median of maximum coherence between locus coeruleus or reference BOLD and HRV are provided.

https://cdn.elifesciences.org/articles/52059/elife-52059-fig5-data1-v2.rdata

Download elife-52059-fig5-data1-v2.rdata

We ran linear mixed effect models to examine whether the relationship between frequency and coherence varied across task stages (Supplementary file 7A), followed by simple slopes analyses to determine the region of significant (RoS) moderation at α <0.05. Greater LC-HRV coherence levels were observed during baseline compared to consolidation (β = −0.20, t(7889) = −6.65, p<0.001, RoS:<0.27 Hz), during encoding compared to consolidation (β = 0.13, t(7889) = 4.31, p<0.001, RoS:<0.29 Hz), during recollection compared to baseline (β = −0.20, t(7889) = −6.55, p<0.001, RoS:>0.25 Hz) and during recollection compared to encoding (β = −0.12, t(7889) = −4.17, p<0.001, RoS:>0.27 Hz; Figure 5C).

For the control region (Figure 5B and D, Supplementary file 7B), coherence between the reference region and HRV coherence was greater during recollection compared to baseline (β = −0.13, t(7889) = −4.15, p=0.001, RoS:>0.28 Hz), or at trend-level compared to consolidation (β = −0.09, t(7889) = −2.95, p=0.050, RoS:>0.33 Hz). These findings were replicated across preprocessing pipelines, except for the spatially normalized pipeline where associations between frequency and coherence were noisier (Figure 5—figure supplements 1–2, Supplementary file 4–5). These patterns indicate greater LC activity coupled to greater parasympathetic inhibition during baseline and recollection, which has been associated with better discrimination between emotional versus neutral faces, attentional resources and better retrieval of context information from stored representation to perform memory task successfully (Lane et al., 2009; Thayer et al., 2012). On the other hand, consolidation demonstrated lower coherence compared to encoding or baseline, which is consistent with the premise that lower HF-HRV is associated with higher arousal. The latter is consistent with the increase in ΔsAA during consolidation (Figure 3B).

To further probe the importance of coupling between autonomic tone and LC variability, we defined the frequency band of maximum coherence as the confidence interval of the median of the maximum-individual coherence between LC and HRV (median = 0.217, 95% CI=[0.189, 0.317]. Interestingly, the width of this CI also nicely fits with the standard definition of high-frequency band of parasympathetic influence on HRV [0.15, 0.4 Hz]). The median coherence within this band was extracted for each individual across conditions to obtain a single point-estimate for our repeated measures correlation analyses with ΔsAA. We observed a negative relationship between LC-HRV median coherence and ΔsAA (R_rm(df)=−0.383 (34), p-value=0.021, CI (95%)=[−0.64,–0.05], Figure 5E). We replicated this analysis across preprocessing stages. Additionally, we repeated this analysis for the reference component, but observed no significant relationship with ΔsAA (R_rm(df)=0.218 (34), p-value=0.201, CI (95%)=[−0.13, 0.52], Figure 5F). Finally, no significant correlation was detected between LC-HRV coherence during any of the stages and emotional recollection rate (all ps > 0.05). The covariation between LC activity and HRV demonstrate that even though both emotion and cognition operate simultaneously, encoding and recollection are coupled to moment-to-moment autonomic regulation supporting cognitive functions required for the task at hand. However, this LC-parasympathetic coupling, appertaining to the central autonomic network, is not sufficient to predict accurate emotional memory recollection. Conversely, LC BOLD signal during consolidation may be specifically tied to arousal-related sympathetic tone, an inference supported by our ΔsAA analyses and the at-trend correlation between ΔsAA during consolidation and memory performance. We should note that the physiological relevance of high-frequency fMRI signal remains unclear, but it has indeed been linked to rapid modulation of the brainstem driving information exchange (Billings et al., 2018).

Arousal-related successful encoding elicits LC activation

We hypothesized that the LC and amygdala would be activated during encoding of emotional compared to neutral faces. The GLM within our predefined mask (Figure 6—figure supplement 1) largely confirmed this, by showing activation in the CA1 and 3, bilateral BLA (extending into the centromedial amygdala) and right entorhinal cortex during successful encoding versus no successful encoding (Supplementary file 8A, Figure 6A). However, no significant clusters were detected for emotional >neutral or for emotional successful encoding >neutral successful encoding.

Figure 6 with 3 supplements see all

Download asset Open asset

Figure 6—source data 1

Locus coeruleus and medial temporal lobe activity during the task stages.

Individual beta’s are provided for the amygdala, hippocampus and locus coeruleus for the examined relevant contrasts during encoding.

https://cdn.elifesciences.org/articles/52059/elife-52059-fig6-data1-v2.rdata

Download elife-52059-fig6-data1-v2.rdata

Given our evidence of the LC’s role in modulation of the autonomic system, we examined contrasts from HRV-convolved clusters. No significant clusters were observed after cluster correction in the MTL or pons. But given the limited signal-to-noise ratio typical for brainstem responses and the small-elongated shape of the LC (10–12 voxels in our high-resolution images), we performed our analyses at an uncorrected p<0.05 within a small-volume search space. The high arousal (local HRV minima) events (High Arousal > Low Arousal) showed significant bilateral activation at the LC (peak z-value = 3.3, Figure 6C), consistent with our previous results examining HRV-LC coherence during encoding. In addition, significant voxels of LC activation were observed for the interaction of high compared to low arousal during successful encoding (High Arousal successful Encoding > Low Arousal successful Encoding; peak z-value = 5.2; Figure 6B), successfully encoding versus forgotten trials during high arousal (High Arousal successful Encoding > High Arousal forgotten, peak z-value = 2.9, Figure 6D) and bilaterally for encoding of emotional stimuli relative to neutral during high arousal (High Arousal Emotional Encoding > High Arousal Neutral Encoding, peak z-value = 5.8, Figure 6E).

The extracted beta-coefficients of this contrast correlated with emotional memory performance (r = 0.506, N = 24, p=0.019), indicating that higher LC activity in emotionally encoded trials under high arousal as compared to neutral trials was related to better memory performance. These results suggest that the MTL and BLA is involved in the accuracy of learning, while LC clusters are involved in encoding of emotional stimuli, a role that is particularly modulated by arousal.

Interactions between the LC and MTL transition across task stages

To compare interactions between regions across the task stages we performed seed-to-voxel analyses using cross-correlation at lag = 0 s (denoted as functional connectivity FC) between the LC using the component of the baseline resting-state (Figure 4—figure supplement 1) and the predefined MTL regions (Figure 6—figure supplement 1): hippocampus (HIPP), amygdala (AMY) and entorhinal cortex (EC). We observed a decrease in FC during consolidation compared to baseline between the LC and left BLA, (para)subiculum and EC (Supplementary file 8C, Figure 7A). No effects were observed when comparing baseline with encoding or recollection and these findings were reproduced across preprocessing pipelines (Figure 7—figure supplement 1).

Figure 7 with 1 supplement see all

Download asset Open asset

Figure 7—source data 1

Coherence and phase between time series of the locus coeruleus and medial temporal lobe clusters.

Individual coherence and phase values between the locus coeruleus and each cluster in the medial temporal lobe (amygdala, hippocampus, entorhinal cortex) are provided per task stage.

https://cdn.elifesciences.org/articles/52059/elife-52059-fig7-data1-v2.rdata

Download elife-52059-fig7-data1-v2.rdata

Areas related to arousal and salience processing, such as the LC, have particularly variable behavior that can result in a negative FC value, complicating their interpretation (Chang and Glover, 2010). As our FC analyses assume stationarity within the fMRI timeseries, we aimed to understand these negative FC changes by exploring the frequency-specific magnitude and phase correlation changes in the follow-up spectral coherence analyses. Phase or magnitude changes in the frequency plane can result in reduced cross-correlation values.

To understand whether timeseries from specific regions were leading or lagging during the task stages, we estimated the median phase lag and coherence across the frequency band of maximum coherence between the timeseries of the LC and MTL regions. Similar to previously determined LC-HRV median maximum coherence (Figure 5), we calculated the confidence interval of the median of the maximum-individual coherence between timeseries. Linear mixed effects models were implemented with either median coherence or phase at maximum coherence between the task changes as outcome and task stage, timeseries combination and their interaction as fixed effects. We observed an increase in HIP-AMY coherence for encoding compared to baseline (β = −0.13, t(821)=-4.39, p<0.001, 95% CI[−0.20,–0.052]), a decrease in recollection compared to encoding (β = 0.10, t(821)=3.37, p=0.004, 95% CI[0.023, 0.171]) and a trend for consolidation decrease compared to baseline (β = −0.07, t(821)=-2.35, p=0.088 95% CI[−0.141, 0.007]) and a change in phase from AMY leading the LC in encoding to the LC leading the AMY in recollection (β = 0.91, t(821)=2.73, p=0.033, 95% CI[0.051,1.772] Figure 7B). Coherent behavior between timeseries may be influenced by other brain regions. Therefore, we calculated partial coherences and compared these against a null distribution consisting of partial coherences from randomized time-series with matched auto-regressive coefficients. The surviving significant coherent relationships between timeseries are visualized in Figure 7 (B–E) in terms of coherence magnitude and phase. Accounting for timeseries of the AMY reduced HIPP-LC coherence during encoding, and accounting for HIPP timeseries significantly reduced coherence between AMY and LC during consolidation.

When relating emotional memory to coherence for these ROI pairs using linear regressions (adjusted for baseline coherence), we observed that greater memory performance was associated with greater AMY-EC coherence during encoding (β = 0.28, t(21)=2.62, p=0.038, 95% CI[0.057,0.508]), and EC-HIP coherence during encoding (at trend level: β = 0.27, t(21)=2.38, p=0.054, 95% CI[0.033,0.512]).

LC structure is associated with sAA and memory performance

Previous studies related LC structure (intensity) to cognitive and autonomic tone measures (Hämmerer et al., 2018; Mather et al., 2017). The LC segmentation is depicted in Figure 8A–B (see also Figure 8—figure supplement 1). The LC TFL intensity normalized to a pontine tegmentum ROI showed a positive correlation with emotional recollection scores (r = 0.56, p=0.004, Figure 8D). Increased motion correlated borderline with TFL intensity ratio (r = 0.38, p=0.071) and emotional recollection (r = 0.34, p=0.108), but adjusting for motion did not alter the positive relationship between normalized TFL intensity and emotional recollection (p=0.004). When subdividing the LC into three equisized parts (Figure 8C), we observed a positive correlation with emotional recollection (Figure 8D) for the medial (r = 0.52, p=0.009) and caudal (r = 0.43, p=0.034) parts of the LC, not rostral (r = 0.37, p=0.074).

Figure 8 with 1 supplement see all

Download asset Open asset

Figure 8—source data 1

Structural locus coeruleus intensity and memory performance.

Individual locus coeruleus intensity values (entire locus coeruleus, rostral, middle and caudal) and emotional memory scores are provided. The emotional memory scores are calculated as the proportion of adjusted emotional scores versus neutral scores as indicated in the methods.

https://cdn.elifesciences.org/articles/52059/elife-52059-fig8-data1-v2.rdata

Download elife-52059-fig8-data1-v2.rdata

We then related the mean TFL intensity ratio to our measure of autonomic tone and observed a positive relation with ΔsAA during consolidation (r = 602, p=0.008, N = 18). When examining the different sections of the LC, this positive relationship was only observed for the caudal part of the LC (r = 0.79, p<0.001, N = 18). No association was observed between TFL intensity and rMSSD or LC BOLD variability.

The LC is connected with evolutionary ancient brainstem regions involved in autonomic functions, but also with the associative neocortical regions, allowing it to be ‘the cognitive limb of a globally conceived sympathetic nervous system’ (Aston-Jones et al., 1991). We set out to investigate the contribution of the LC and its interactions with MTL regions during arousal-related memory enhancement by combining dedicated ultra-high-field MRI methods, an established emotional memory task, physiological and saliva measures. Consistent with previous work, we observed that forming adaptive memories in response to arousing events was predominantly associated with NE- activity. Considerably less is known about the LC’s interactions with MTL regions during the different memory processing stages. Our results now demonstrate that LC’s role during these memory processing stages was shaped by arousal-related autonomic changes and its close interactions with the amygdala. During encoding, stimuli eliciting arousal-related autonomic changes were associated with LC activity following amygdala activation during successful learning. During consolidation, the role of the LC gradually changed from being arousal-driven to maintenance or updating internal models during recollection, involving predominantly hippocampal activity. These results highlight the dynamic nature of neuromodulatory brain networks during arousal-related memory processes (Hermans et al., 2011). Understanding the neural mechanisms underlying memory for emotional experiences increases our understanding of diseases in which a dysregulation or decoupling of one of these systems could lead to affective, autonomic or neurocognitive disorders (Samuels and Szabadi, 2008b; Jacobs et al., 2019).

Activation of the LC occurs in parallel with the autonomic system, and modulates cognition: across the entire memory task greater variability in LC BOLD signal was associated with lower parasympathetic activity (rMSSD), which in turn was associated with greater sympathetic modulation (ΔsAA) and better memory performance. Beyond function, we also observed that structural properties of the LC, the middle section, related to better emotional memory performance, and greater caudal LC intensity was related to greater NE. These correlations fit with the well-described topographic projections of the LC, where the rostral and middle part project densely to higher-order associative regions of the brain and the caudal part of the LC to the nucleus ambiguus and dorsal motor nucleus of the vagus (Sara, 2009; Samuels and Szabadi, 2008a), which both modulate the parasympathetic outflow based on sympathetic signals.

We also observed task-stage-specific associations between LC’s connectivity to the MTL regions and autonomic tone measures. During consolidation, we observed a different LC-MTL functional connectivity coupling relative to baseline, possibly reflecting transient processes related to less parasympathetic tone (lower coherence in the high frequency domain). Correspondingly, increases in sAA during consolidation correlated positively with emotional performance. Such increases in sympathetic tone may reflect arousal-related NE-synaptic consolidation processes (Packard et al., 1994).

While systems consolidation requires years, synaptic consolidation represents the initial phase after learning where mnemonic representations change and new memory tracers are formed. Previous reports demonstrated that synaptic consolidation may be captured by the resting-state signal after learning a task (Vilberg and Davachi, 2013; Tambini et al., 2010; Jacobs et al., 2014). This consolidation phase may be a transitional stage, where the role of the LC slowly alters from an arousal-based learning experience to maintenance of memories via long-term potentiation (LTP). LTP depends on the activation of the BLA, highlighting the critical role of the amygdala (Frey et al., 2001). A myriad of animal studies have demonstrated that elevated NE in the amygdala affects memory consolidation for stress-induced events and can facilitate hippocampal synaptic plasticity (Hermans et al., 2014; McIntyre et al., 2003).

In support of this premise, we observed mainly arousal-related LC activity during encoding. During encoding, the LC was activated during (successful) acquisition of information that was associated with elevated arousal. However, amygdala and hippocampal activation was only detected for successful learning, independent of valence of arousal. Arousal-related amygdala activation may have been attenuated by the cognitive task during encoding (Lange et al., 2003). Nonetheless, arousal-related LC-activation with neural input from amygdala-entorhinal co-activation was associated with memory enhancement. This is largely consistent with work reporting that sympathetic arousal can predict memory performance for high-arousing stimuli when this coincides with amygdala activation (de Voogd et al., 2016). During recollection, we observed no arousal-related activity, but successful recollection was associated with CA1 activity. Activation of β-adrenoreceptors in the CA1 during consolidation is required for later retention of the memory task, but not for the aversive component of the task (Mello-Carpes and Izquierdo, 2013; Chen et al., 1992). The markedly higher frequency window of the HRV-LC coherences during recollection, suggests less NE-sympathetic responsiveness at this stage. Increases in LC-hippocampus and LC-amygdala coherence variability during recollection are possibly related to increased parasympathetic inhibitory modulation supporting cognitive operations. Alternatively, it could be speculated that spill-over from NE- sympathetic processes during consolidation drive retention processes.

In addition to different activation patterns across stages, we also observed a switch in the role of LC’s connectivity with MTL regions. During encoding, the LC is on the receiving end, in particular from the amygdala, whereas this changes to a leading role in recollection. This phase switch between the LC and amygdala conforms with functional and anatomical descriptions of projections between the LC and central amygdala or BLA. The LC receives afferents from the central amygdala mainly at the periLC-region (Bergado et al., 2007). The central amygdala controls stress responses and is involved in salience processing and attention processes, important for efficient encoding (Phelps and LeDoux, 2005; Roozendaal et al., 1996). While the central amygdala has no effect on hippocampal LTP, the BLA has memory-enhancing effects on the subiculum and CA1 via β-adrenergic mechanisms (Phelps and LeDoux, 2005; Roozendaal et al., 1996). The NE-input to the BLA arrives mainly from the LC (Chen and Sara, 2007). The BLA does not project directly to the LC, but via a small direct projection to the central amygdala (Bouret et al., 2003).

Recent general theories of brain function, such as the active inference theory (Friston et al., 2015), posit that NE responses from the LC can be considered as a prediction error during learning originating from information from the amygdala, whereas signals from the LC to the amygdala and cortex guide transfer of information and model updating (Sales et al., 2019). In these models, the amygdala is critical to maintain allostasis. The central amygdala network would control the arousing physiological responses through its descending predicting signals to the brainstem, whereas the BLA would be part of a neutral signal uncertainty network as it receives prediction errors from the LC. It has been suggested that NE increases accuracy of episodic memories (Barsegyan et al., 2014). While our study was not designed to test the active inference theory, the proposed models are consistent with our phase-differences across stages. Future in-vivo recordings, computational and carefully designed studies are warranted to test these models in detail.

No study is without limitations. First, as our focus was on the LC and MTL, we specifically opted to acquire fMR images with a high isotropic spatial resolution. This limited our field of view and therefore, we did not investigate distant modulatory effects. Second, the LC is challenging to investigate in humans due to limits in spatiotemporal resolution and lower signal-to-noise ratio in the brainstem compared to neocortical signal. We had to eschew cluster-based correction. The small in-plane diameter of the LC means that only low-levels of smoothing can be applied, which would reduce the signal-to-noise, and the probability of discovering a contiguous cluster in the highly-elongated axial plane of the structure. Ultimately, this would result in an overly stringent multiple comparison adjustment. Nevertheless, we are confident that our GLM findings do not represent false positives due to their reproducibility across preprocessing pipelines, correlation to behavior and consistent anatomical co-localization with an ex-vivo validated template and anatomical atlases.

To conclude, interactions between the NE-LC, amygdala and hippocampus, drive arousal-related memory processing, but transitions in interactions occur along the memory processing stages. During successful encoding, we observed high-arousal-related LC activation, possibly reflecting attentional processes, in conjunction with amygdala activation. During consolidation, LC-MTL connectivity is dynamic and associated with NE increases possibly reflecting LTP. This gradually transitions to learning and updating internal models during recollection, which mainly involves hippocampal activity. These findings contribute to our understanding of LC activation for forming adaptive memories of emotional experiences and can serve as a basis for understanding how LC dysregulation may potentially disrupt this adaptive behavior.

Processed data to reproduce the figures in this manuscript is provided in the Source data files. Participants did not explicitly consent to their data being made public and therefore, access to their demographic, raw or processed imaging and physiological data is restricted. Requests for the anonymized data should be made to Heidi Jacobs (www.heidijacobs.nl; h.jacobs@maastrichtuniversity.nl or hjacobs@mgh.harvard.edu) and will be reviewed by an independent data access committee, taking into account the research proposal and intended use of the data. Requestors are required to sign a data sharing agreement to ensure participants' confidentiality is maintained prior to release of any data and that procedures conform with the EU legislation on the general data protection regulation and local ethical regulations.

1. 1. Akselrod S
   2. Gordon D
   3. Ubel FA
   4. Shannon DC
   5. Berger AC
   6. Cohen RJ

   (1981) Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat-to-beat cardiovascular control

   Science 213:220–222.

   https://doi.org/10.1126/science.6166045

   • PubMed
   • Google Scholar
2. 1. Aston-Jones G
   2. Shipley MT
   3. Chouvet G
   4. Ennis M
   5. van Bockstaele E
   6. Pieribone V
   7. Shiekhattar R
   8. Akaoka H
   9. Drolet G
   10. Astier B

   (1991) Afferent regulation of locus coeruleus neurons: anatomy, physiology and pharmacology

   Progress in Brain Research 88:47–75.

   https://doi.org/10.1016/s0079-6123(08)63799-1

   • PubMed
   • Google Scholar
3. 1. Avants BB
   2. Tustison NJ
   3. Song G
   4. Cook PA
   5. Klein A
   6. Gee JC

   (2011) A reproducible evaluation of ANTs similarity metric performance in brain image registration

   NeuroImage 54:2033–2044.

   https://doi.org/10.1016/j.neuroimage.2010.09.025

   • PubMed
   • Google Scholar
4. 1. Barsegyan A
   2. McGaugh JL
   3. Roozendaal B

   (2014) Noradrenergic activation of the basolateral amygdala modulates the consolidation of object-in-context recognition memory

   Frontiers in Behavioral Neuroscience 8:160.

   https://doi.org/10.3389/fnbeh.2014.00160

   • PubMed
   • Google Scholar
5. 1. Bergado JA
   2. Frey S
   3. López J
   4. Almaguer-Melian W
   5. Frey JU

   (2007) Cholinergic afferents to the locus coeruleus and noradrenergic afferents to the medial septum mediate LTP-reinforcement in the dentate gyrus by stimulation of the amygdala

   Neurobiology of Learning and Memory 88:331–341.

   https://doi.org/10.1016/j.nlm.2007.05.003

   • PubMed
   • Google Scholar
6. 1. Billings JCW
   2. Thompson GJ
   3. Pan WJ
   4. Magnuson ME
   5. Medda A
   6. Keilholz S

   (2018) Disentangling multispectral functional connectivity with wavelets

   Frontiers in Neuroscience 12:812.

   https://doi.org/10.3389/fnins.2018.00812

   • PubMed
   • Google Scholar
7. 1. Bittiner SB
   2. Smith SE

   (1986) Beta-adrenoceptor antagonists increase sinus arrhythmia, a vagotonic effect

   British Journal of Clinical Pharmacology 22:691–695.

   https://doi.org/10.1111/j.1365-2125.1986.tb02959.x

   • PubMed
   • Google Scholar
8. 1. Bouret S
   2. Duvel A
   3. Onat S
   4. Sara SJ

   (2003) Phasic activation of locus ceruleus neurons by the central nucleus of the amygdala

   The Journal of Neuroscience 23:3491–3497.

   https://doi.org/10.1523/JNEUROSCI.23-08-03491.2003

   • PubMed
   • Google Scholar
9. 1. Bouret S
   2. Sara SJ

   (2005) Network reset: a simplified overarching theory of locus coeruleus noradrenaline function

   Trends in Neurosciences 28:574–582.

   https://doi.org/10.1016/j.tins.2005.09.002

   • PubMed
   • Google Scholar
10. 1. Cahill L
    2. Prins B
    3. Weber M
    4. McGaugh JL

    (1994) Beta-adrenergic activation and memory for emotional events

    Nature 371:702–704.

    https://doi.org/10.1038/371702a0

    • PubMed
    • Google Scholar
11. 1. Chang C
    2. Glover GH

    (2010) Time-frequency dynamics of resting-state brain connectivity measured with fMRI

    NeuroImage 50:81–98.

    https://doi.org/10.1016/j.neuroimage.2009.12.011

    • PubMed
    • Google Scholar
12. 1. Chen MF
    2. Chiu TH
    3. Lee EH

    (1992) Noradrenergic mediation of the memory-enhancing effect of corticotropin-releasing factor in the locus coeruleus of rats

    Psychoneuroendocrinology 17:113–124.

    https://doi.org/10.1016/0306-4530(92)90050-H

    • PubMed
    • Google Scholar
13. 1. Chen FJ
    2. Sara SJ

    (2007) Locus coeruleus activation by foot shock or electrical stimulation inhibits amygdala neurons

    Neuroscience 144:472–481.

    https://doi.org/10.1016/j.neuroscience.2006.09.037

    • PubMed
    • Google Scholar
14. 1. Dale AM

    (1999) Optimal experimental design for event-related fMRI

    Human Brain Mapping 8:109–114.

    https://doi.org/10.1002/(SICI)1097-0193(1999)8:2/3<109::AID-HBM7>3.0.CO;2-W

    • PubMed
    • Google Scholar
15. 1. de Voogd LD
    2. Fernández G
    3. Hermans EJ

    (2016) Disentangling the roles of arousal and amygdala activation in emotional declarative memory

    Social Cognitive and Affective Neuroscience 11:1471–1480.

    https://doi.org/10.1093/scan/nsw055

    • PubMed
    • Google Scholar
16. 1. Ekman P

    (1992) Facial expressions of emotion: an old controversy and new findings

    Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 335:63–69.

    https://doi.org/10.1098/rstb.1992.0008

    • PubMed
    • Google Scholar
17. 1. Fischl B
    2. Sereno MI
    3. Dale AM

    (1999) Cortical surface-based analysis II: inflation, flattening, and a surface-based coordinate system

    NeuroImage 9:195–207.

    https://doi.org/10.1006/nimg.1998.0396

    • PubMed
    • Google Scholar
18. 1. Frey S
    2. Bergado-Rosado J
    3. Seidenbecher T
    4. Pape HC
    5. Frey JU

    (2001) Reinforcement of early long-term potentiation (early-LTP) in Dentate Gyrus by stimulation of the basolateral amygdala: heterosynaptic induction mechanisms of late-LTP

    The Journal of Neuroscience 21:3697–3703.

    https://doi.org/10.1523/JNEUROSCI.21-10-03697.2001

    • PubMed
    • Google Scholar
19. 1. Friston K
    2. Rigoli F
    3. Ognibene D
    4. Mathys C
    5. Fitzgerald T
    6. Pezzulo G

    (2015) Active inference and epistemic value

    Cognitive Neuroscience 6:187–214.

    https://doi.org/10.1080/17588928.2015.1020053

    • PubMed
    • Google Scholar
20. 1. Griffanti L
    2. Salimi-Khorshidi G
    3. Beckmann CF
    4. Auerbach EJ
    5. Douaud G
    6. Sexton CE
    7. Zsoldos E
    8. Ebmeier KP
    9. Filippini N
    10. Mackay CE
    11. Moeller S
    12. Xu J
    13. Yacoub E
    14. Baselli G
    15. Ugurbil K
    16. Miller KL
    17. Smith SM

    (2014) ICA-based artefact removal and accelerated fMRI acquisition for improved resting state network imaging

    NeuroImage 95:232–247.

    https://doi.org/10.1016/j.neuroimage.2014.03.034

    • PubMed
    • Google Scholar
21. 1. Gur RC
    2. Sara R
    3. Hagendoorn M
    4. Marom O
    5. Hughett P
    6. Macy L
    7. Turner T
    8. Bajcsy R
    9. Posner A
    10. Gur RE

    (2002) A method for obtaining 3-dimensional facial expressions and its standardization for use in neurocognitive studies

    Journal of Neuroscience Methods 115:137–143.

    https://doi.org/10.1016/S0165-0270(02)00006-7

    • PubMed
    • Google Scholar
22. 1. Hämmerer D
    2. Callaghan MF
    3. Hopkins A
    4. Kosciessa J
    5. Betts M
    6. Cardenas-Blanco A
    7. Kanowski M
    8. Weiskopf N
    9. Dayan P
    10. Dolan RJ
    11. Düzel E

    (2018) Locus coeruleus integrity in old age is selectively related to memories linked with salient negative events

    PNAS 115:2228–2233.

    https://doi.org/10.1073/pnas.1712268115

    • PubMed
    • Google Scholar
23. 1. Hansen N
    2. Manahan-Vaughan D

    (2015) Locus coeruleus stimulation facilitates Long-Term depression in the dentate gyrus that requires activation of -Adrenergic Receptors

    Cerebral Cortex 25:1889–1896.

    https://doi.org/10.1093/cercor/bht429

    • Google Scholar
24. 1. Harvey AK
    2. Pattinson KT
    3. Brooks JC
    4. Mayhew SD
    5. Jenkinson M
    6. Wise RG

    (2008) Brainstem functional magnetic resonance imaging: disentangling signal from physiological noise

    Journal of Magnetic Resonance Imaging 28:1337–1344.

    https://doi.org/10.1002/jmri.21623

    • PubMed
    • Google Scholar
25. 1. Hedman AE
    2. Hartikainen JE
    3. Tahvanainen KU
    4. Hakumäki MO

    (1995) The high frequency component of heart rate variability reflects cardiac parasympathetic modulation rather than parasympathetic 'tone'

    Acta Physiologica Scandinavica 155:267–273.

    https://doi.org/10.1111/j.1748-1716.1995.tb09973.x

    • PubMed
    • Google Scholar
26. 1. Hermans EJ
    2. van Marle HJ
    3. Ossewaarde L
    4. Henckens MJ
    5. Qin S
    6. van Kesteren MT
    7. Schoots VC
    8. Cousijn H
    9. Rijpkema M
    10. Oostenveld R
    11. Fernández G

    (2011) Stress-related noradrenergic activity prompts large-scale neural network reconfiguration

    Science 334:1151–1153.

    https://doi.org/10.1126/science.1209603

    • PubMed
    • Google Scholar
27. 1. Hermans EJ
    2. Battaglia FP
    3. Atsak P
    4. de Voogd LD
    5. Fernández G
    6. Roozendaal B

    (2014) How the amygdala affects emotional memory by altering brain network properties

    Neurobiology of Learning and Memory 112:2–16.

    https://doi.org/10.1016/j.nlm.2014.02.005

    • PubMed
    • Google Scholar
28. 1. Hill LK
    2. Siebenbrock A

    (2009)

    Are all measures created equal? heart rate variability and respiration - biomed 2009

    Biomedical Sciences Instrumentation 45:71–76.

    • PubMed
    • Google Scholar
29. 1. Iglesias JE
    2. Augustinack JC
    3. Nguyen K
    4. Player CM
    5. Player A
    6. Wright M
    7. Roy N
    8. Frosch MP
    9. McKee AC
    10. Wald LL
    11. Fischl B
    12. Van Leemput K
    13. Alzheimer's Disease Neuroimaging Initiative

    (2015a) A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: application to adaptive segmentation of in vivo MRI

    NeuroImage 115:117–137.

    https://doi.org/10.1016/j.neuroimage.2015.04.042

    • PubMed
    • Google Scholar
30. 1. Iglesias JE
    2. Van Leemput K
    3. Bhatt P
    4. Casillas C
    5. Dutt S
    6. Schuff N
    7. Truran-Sacrey D
    8. Boxer A
    9. Fischl B
    10. Alzheimer's Disease Neuroimaging Initiative

    (2015b) Bayesian segmentation of brainstem structures in MRI

    NeuroImage 113:184–195.

    https://doi.org/10.1016/j.neuroimage.2015.02.065

    • PubMed
    • Google Scholar
31. 1. Jacobs HI
    2. Dillen KN
    3. Risius O
    4. Göreci Y
    5. Onur OA
    6. Fink GR
    7. Kukolja J

    (2014) Consolidation in older adults depends upon competition between resting-state networks

    Frontiers in Aging Neuroscience 6:344.

    https://doi.org/10.3389/fnagi.2014.00344

    • PubMed
    • Google Scholar
32. 1. Jacobs HI
    2. Riphagen JM
    3. Razat CM
    4. Wiese S
    5. Sack AT

    (2015) Transcutaneous vagus nerve stimulation boosts associative memory in older individuals

    Neurobiology of Aging 36:1860–1867.

    https://doi.org/10.1016/j.neurobiolaging.2015.02.023

    • PubMed
    • Google Scholar
33. 1. Jacobs HIL
    2. Riphagen JM
    3. Ramakers I
    4. Verhey FRJ

    (2019) Alzheimer’s disease pathology: pathways between central norepinephrine activity, memory, and neuropsychiatric symptoms

    Molecular Psychiatry 82:437.

    https://doi.org/10.1038/s41380-019-0437-x

    • Google Scholar
34. 1. Jenkinson M
    2. Bannister P
    3. Brady M
    4. Smith S

    (2002) Improved optimization for the robust and accurate linear registration and motion correction of brain images

    NeuroImage 17:825–841.

    https://doi.org/10.1006/nimg.2002.1132

    • PubMed
    • Google Scholar
35. 1. Keren NI
    2. Lozar CT
    3. Harris KC
    4. Morgan PS
    5. Eckert MA

    (2009) In vivo mapping of the human locus coeruleus

    NeuroImage 47:1261–1267.

    https://doi.org/10.1016/j.neuroimage.2009.06.012

    • PubMed
    • Google Scholar
36. 1. Khakpour-Taleghani B
    2. Lashgari R
    3. Aavani T
    4. Haghparast A
    5. Naderi N
    6. Motamedi F

    (2008) The locus coeruleus involves in consolidation and memory retrieval, but not in acquisition of inhibitory avoidance learning task

    Behavioural Brain Research 189:257–262.

    https://doi.org/10.1016/j.bbr.2008.01.004

    • PubMed
    • Google Scholar
37. 1. Kleiger RE
    2. Stein PK
    3. Bigger JT

    (2005) Heart rate variability: measurement and clinical utility

    Annals of Noninvasive Electrocardiology 10:88–101.

    https://doi.org/10.1111/j.1542-474X.2005.10101.x

    • PubMed
    • Google Scholar
38. 1. Klein A
    2. Andersson J
    3. Ardekani BA
    4. Ashburner J
    5. Avants B
    6. Chiang MC
    7. Christensen GE
    8. Collins DL
    9. Gee J
    10. Hellier P
    11. Song JH
    12. Jenkinson M
    13. Lepage C
    14. Rueckert D
    15. Thompson P
    16. Vercauteren T
    17. Woods RP
    18. Mann JJ
    19. Parsey RV

    (2009) Evaluation of 14 nonlinear deformation algorithms applied to human brain MRI registration

    NeuroImage 46:786–802.

    https://doi.org/10.1016/j.neuroimage.2008.12.037

    • PubMed
    • Google Scholar
39. 1. Lane RD
    2. McRae K
    3. Reiman EM
    4. Chen K
    5. Ahern GL
    6. Thayer JF

    (2009) Neural correlates of heart rate variability during emotion

    NeuroImage 44:213–222.

    https://doi.org/10.1016/j.neuroimage.2008.07.056

    • PubMed
    • Google Scholar
40. 1. Lange K
    2. Williams LM
    3. Young AW
    4. Bullmore ET
    5. Brammer MJ
    6. Williams SC
    7. Gray JA
    8. Phillips ML

    (2003) Task instructions modulate neural responses to fearful facial expressions

    Biological Psychiatry 53:226–232.

    https://doi.org/10.1016/S0006-3223(02)01455-5

    • PubMed
    • Google Scholar
41. 1. Langner O
    2. Dotsch R
    3. Bijlstra G
    4. Wigboldus DHJ
    5. Hawk ST
    6. van Knippenberg A

    (2010) Presentation and validation of the radboud faces database

    Cognition & Emotion 24:1377–1388.

    https://doi.org/10.1080/02699930903485076

    • Google Scholar
42. 1. Lemon N
    2. Aydin-Abidin S
    3. Funke K
    4. Manahan-Vaughan D

    (2009) Locus coeruleus activation facilitates memory encoding and induces hippocampal LTD that depends on beta-adrenergic receptor activation

    Cerebral Cortex 19:2827–2837.

    https://doi.org/10.1093/cercor/bhp065

    • PubMed
    • Google Scholar
43. 1. Marques JP
    2. Kober T
    3. Krueger G
    4. van der Zwaag W
    5. Van de Moortele PF
    6. Gruetter R

    (2010) MP2RAGE, a self bias-field corrected sequence for improved segmentation and T1-mapping at high field

    NeuroImage 49:1271–1281.

    https://doi.org/10.1016/j.neuroimage.2009.10.002

    • PubMed
    • Google Scholar
44. 1. Mather M
    2. Clewett D
    3. Sakaki M
    4. Harley CW

    (2016) Norepinephrine ignites local hotspots of neuronal excitation: how arousal amplifies selectivity in perception and memory

    Behavioral and Brain Sciences 39:e200.

    https://doi.org/10.1017/S0140525X15000667

    • PubMed
    • Google Scholar
45. 1. Mather M
    2. Joo Yoo H
    3. Clewett DV
    4. Lee TH
    5. Greening SG
    6. Ponzio A
    7. Min J
    8. Thayer JF

    (2017) Higher locus coeruleus MRI contrast is associated with lower parasympathetic influence over heart rate variability

    NeuroImage 150:329–335.

    https://doi.org/10.1016/j.neuroimage.2017.02.025

    • PubMed
    • Google Scholar
46. 1. McIntyre CK
    2. Power AE
    3. Roozendaal B
    4. McGaugh JL

    (2003) Role of the basolateral amygdala in memory consolidation

    Annals of the New York Academy of Sciences 985:273–293.

    https://doi.org/10.1111/j.1749-6632.2003.tb07088.x

    • PubMed
    • Google Scholar
47. 1. Mello-Carpes PB
    2. Izquierdo I

    (2013) The nucleus of the solitary tract → nucleus paragigantocellularis → locus coeruleus → CA1 region of dorsal Hippocampus pathway is important for consolidation of object recognition memory

    Neurobiology of Learning and Memory 100:56–63.

    https://doi.org/10.1016/j.nlm.2012.12.002

    • PubMed
    • Google Scholar
48. 1. Moeller S
    2. Yacoub E
    3. Olman CA
    4. Auerbach E
    5. Strupp J
    6. Harel N
    7. Uğurbil K

    (2010) Multiband multislice GE-EPI at 7 tesla, with 16-fold acceleration using partial parallel imaging with application to high spatial and temporal whole-brain fMRI

    Magnetic Resonance in Medicine 63:1144–1153.

    https://doi.org/10.1002/mrm.22361

    • PubMed
    • Google Scholar
49. 1. Nater UM
    2. Rohleder N

    (2009) Salivary alpha-amylase as a non-invasive biomarker for the sympathetic nervous system: current state of research

    Psychoneuroendocrinology 34:486–496.

    https://doi.org/10.1016/j.psyneuen.2009.01.014

    • PubMed
    • Google Scholar
50. 1. Packard MG
    2. Cahill L
    3. McGaugh JL

    (1994) Amygdala modulation of hippocampal-dependent and caudate nucleus-dependent memory processes

    PNAS 91:8477–8481.

    https://doi.org/10.1073/pnas.91.18.8477

    • PubMed
    • Google Scholar
51. 1. Phelps EA
    2. LeDoux JE

    (2005) Contributions of the amygdala to emotion processing: from animal models to human behavior

    Neuron 48:175–187.

    https://doi.org/10.1016/j.neuron.2005.09.025

    • PubMed
    • Google Scholar
52. Conference

    1. Pinheiro N
    2. Couceiro R
    3. Henriques J
    4. Muehlsteff J
    5. Quintal I
    6. Goncalves L
    7. Carvalho P

    (2016) Can PPG be used for HRV analysis?

    Conference Proceedings : . Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference. pp. 2945–2949.

    https://doi.org/10.1109/EMBC.2016.7591347

    • Google Scholar
53. 1. Power AE
    2. Thal LJ
    3. McGaugh JL

    (2002) Lesions of the nucleus basalis magnocellularis induced by 192 IgG-saporin block memory enhancement with posttraining norepinephrine in the basolateral amygdala

    PNAS 99:2315–2319.

    https://doi.org/10.1073/pnas.022627799

    • PubMed
    • Google Scholar
54. 1. Priovoulos N
    2. Jacobs HIL
    3. Ivanov D
    4. Uludağ K
    5. Verhey FRJ
    6. Poser BA

    (2018) High-resolution in vivo imaging of human locus coeruleus by magnetization transfer MRI at 3T and 7T

    NeuroImage 168:427–436.

    https://doi.org/10.1016/j.neuroimage.2017.07.045

    • PubMed
    • Google Scholar
55. 1. Rimmele U
    2. Lackovic SF
    3. Tobe RH
    4. Leventhal BL
    5. Phelps EA

    (2016) Beta-adrenergic blockade at memory encoding, but not retrieval, decreases the subjective sense of recollection

    Journal of Cognitive Neuroscience 28:895–907.

    https://doi.org/10.1162/jocn_a_00941

    • PubMed
    • Google Scholar
56. 1. Rohleder N
    2. Nater UM

    (2009) Determinants of salivary alpha-amylase in humans and methodological considerations

    Psychoneuroendocrinology 34:469–485.

    https://doi.org/10.1016/j.psyneuen.2008.12.004

    • PubMed
    • Google Scholar
57. 1. Roozendaal B
    2. Portillo-Marquez G
    3. McGaugh JL

    (1996) Basolateral amygdala lesions block glucocorticoid-induced modulation of memory for spatial learning

    Behavioral Neuroscience 110:1074–1083.

    https://doi.org/10.1037/0735-7044.110.5.1074

    • PubMed
    • Google Scholar
58. 1. Roozendaal B
    2. Castello NA
    3. Vedana G
    4. Barsegyan A
    5. McGaugh JL

    (2008) Noradrenergic activation of the basolateral amygdala modulates consolidation of object recognition memory

    Neurobiology of Learning and Memory 90:576–579.

    https://doi.org/10.1016/j.nlm.2008.06.010

    • PubMed
    • Google Scholar
59. 1. Roozendaal B
    2. Hermans EJ

    (2017) Norepinephrine effects on the encoding and consolidation of emotional memory: improving synergy between animal and human studies

    Current Opinion in Behavioral Sciences 14:115–122.

    https://doi.org/10.1016/j.cobeha.2017.02.001

    • Google Scholar
60. 1. Roozendaal B
    2. McGaugh JL

    (2011) Memory modulation

    Behavioral Neuroscience 125:797–824.

    https://doi.org/10.1037/a0026187

    • PubMed
    • Google Scholar
61. 1. Rugg MD
    2. Vilberg KL

    (2013) Brain networks underlying episodic memory retrieval

    Current Opinion in Neurobiology 23:255–260.

    https://doi.org/10.1016/j.conb.2012.11.005

    • PubMed
    • Google Scholar
62. 1. Sales AC
    2. Friston KJ
    3. Jones MW
    4. Pickering AE
    5. Moran RJ

    (2019) Locus coeruleus tracking of prediction errors optimises cognitive flexibility: an active inference model

    PLOS Computational Biology 15:e1006267.

    https://doi.org/10.1371/journal.pcbi.1006267

    • PubMed
    • Google Scholar
63. 1. Samuels ER
    2. Szabadi E

    (2008a) Functional neuroanatomy of the noradrenergic locus coeruleus: its roles in the regulation of arousal and autonomic function part I: principles of functional organisation

    Current Neuropharmacology 6:235–253.

    https://doi.org/10.2174/157015908785777229

    • PubMed
    • Google Scholar
64. 1. Samuels ER
    2. Szabadi E

    (2008b) Functional neuroanatomy of the noradrenergic locus coeruleus: its roles in the regulation of arousal and autonomic function part II: physiological and pharmacological manipulations and pathological alterations of locus coeruleus activity in humans

    Current Neuropharmacology 6:254–285.

    https://doi.org/10.2174/157015908785777193

    • PubMed
    • Google Scholar
65. 1. Sara SJ

    (2009) The locus coeruleus and noradrenergic modulation of cognition

    Nature Reviews Neuroscience 10:211–223.

    https://doi.org/10.1038/nrn2573

    • PubMed
    • Google Scholar
66. 1. Sara SJ
    2. Bouret S

    (2012) Orienting and reorienting: the locus coeruleus mediates cognition through arousal

    Neuron 76:130–141.

    https://doi.org/10.1016/j.neuron.2012.09.011

    • PubMed
    • Google Scholar
67. 1. Setsompop K
    2. Gagoski BA
    3. Polimeni JR
    4. Witzel T
    5. Wedeen VJ
    6. Wald LL

    (2012) Blipped-controlled aliasing in parallel imaging for simultaneous multislice Echo planar imaging with reduced g-factor penalty

    Magnetic Resonance in Medicine 67:1210–1224.

    https://doi.org/10.1002/mrm.23097

    • PubMed
    • Google Scholar
68. 1. Shaffer F
    2. Ginsberg JP

    (2017) An overview of heart rate variability metrics and norms

    Frontiers in Public Health 5:258.

    https://doi.org/10.3389/fpubh.2017.00258

    • Google Scholar
69. 1. Shmuel A
    2. Yacoub E
    3. Chaimow D
    4. Logothetis NK
    5. Ugurbil K

    (2007) Spatio-temporal point-spread function of fMRI signal in human gray matter at 7 tesla

    NeuroImage 35:539–552.

    https://doi.org/10.1016/j.neuroimage.2006.12.030

    • PubMed
    • Google Scholar
70. 1. Sperling RA
    2. Bates JF
    3. Cocchiarella AJ
    4. Schacter DL
    5. Rosen BR
    6. Albert MS

    (2001) Encoding novel face-name associations: a functional MRI study

    Human Brain Mapping 14:129–139.

    https://doi.org/10.1002/hbm.1047

    • PubMed
    • Google Scholar
71. 1. Stanislaw H
    2. Todorov N

    (1999) Calculation of signal detection theory measures

    Behavior Research Methods, Instruments, & Computers 31:137–149.

    https://doi.org/10.3758/BF03207704

    • Google Scholar
72. 1. Strange BA
    2. Dolan RJ

    (2004) Beta-adrenergic modulation of emotional memory-evoked human amygdala and hippocampal responses

    PNAS 101:11454–11458.

    https://doi.org/10.1073/pnas.0404282101

    • PubMed
    • Google Scholar
73. 1. Tambini A
    2. Ketz N
    3. Davachi L

    (2010) Enhanced brain correlations during rest are related to memory for recent experiences

    Neuron 65:280–290.

    https://doi.org/10.1016/j.neuron.2010.01.001

    • PubMed
    • Google Scholar
74. 1. Thayer JF
    2. Ahs F
    3. Fredrikson M
    4. Sollers JJ
    5. Wager TD

    (2012) A meta-analysis of heart rate variability and neuroimaging studies: implications for heart rate variability as a marker of stress and health

    Neuroscience & Biobehavioral Reviews 36:747–756.

    https://doi.org/10.1016/j.neubiorev.2011.11.009

    • PubMed
    • Google Scholar
75. 1. van Stegeren AH

    (2008) The role of the noradrenergic system in emotional memory

    Acta Psychologica 127:532–541.

    https://doi.org/10.1016/j.actpsy.2007.10.004

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

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

    PNAS 115:E9439–E9448.

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

    • PubMed
    • Google Scholar
77. 1. Vila J
    2. Palacios F
    3. Presedo J
    4. Fernández-Delgado M
    5. Felix P
    6. Barro S

    (1997) Time-frequency analysis of heart-rate variability

    IEEE Engineering in Medicine and Biology Magazine 16:119–126.

    https://doi.org/10.1109/51.620503

    • PubMed
    • Google Scholar
78. 1. Vilberg KL
    2. Davachi L

    (2013) Perirhinal-hippocampal connectivity during reactivation is a marker for object-based memory consolidation

    Neuron 79:1232–1242.

    https://doi.org/10.1016/j.neuron.2013.07.013

    • PubMed
    • Google Scholar
79. 1. Wagatsuma A
    2. Okuyama T
    3. Sun C
    4. Smith LM
    5. Abe K
    6. Tonegawa S

    (2018) Locus coeruleus input to hippocampal CA3 drives single-trial learning of a novel context

    PNAS 115:E310–E316.

    https://doi.org/10.1073/pnas.1714082115

    • PubMed
    • Google Scholar
80. 1. Yasuma F
    2. Hayano J

    (2004) Respiratory sinus arrhythmia: why does the heartbeat synchronize with respiratory rhythm?

    Chest 125:683–690.

    https://doi.org/10.1378/chest.125.2.683

    • PubMed
    • Google Scholar
81. 1. Yonelinas AP
    2. Aly M
    3. Wang WC
    4. Koen JD

    (2010) Recollection and familiarity: examining controversial assumptions and new directions

    Hippocampus 20:1178–1194.

    https://doi.org/10.1002/hipo.20864

    • PubMed
    • Google Scholar
82. 1. Zerbi V
    2. Floriou-Servou A
    3. Markicevic M
    4. Vermeiren Y
    5. Sturman O
    6. Privitera M
    7. von Ziegler L
    8. Ferrari KD
    9. Weber B
    10. De Deyn PP
    11. Wenderoth N
    12. Bohacek J

    (2019) Rapid reconfiguration of the functional connectome after chemogenetic locus coeruleus activation

    Neuron 103:702–718.

    https://doi.org/10.1016/j.neuron.2019.05.034

    • PubMed
    • Google Scholar

Author details

1. Heidi IL Jacobs

   1. Gordon Center for Medical Imaging, Department of Radiology, Massachusetts General Hospital/Harvard Medical School, Boston, United States
   2. Faculty of Health, Medicine and Life Sciences, School for Mental Health and Neuroscience, Alzheimer Centre Limburg, Maastricht University, Maastricht, Netherlands
   3. Faculty of Psychology and Neuroscience, Department of Cognitive Neuroscience, Maastricht University, Maastricht, Netherlands

   Contribution

   Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Project administration, Data interpretation

   Contributed equally with

   Nikos Priovoulos

   For correspondence

   h.jacobs@maastrichtuniversity.nl

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-7620-3822

2. Nikos Priovoulos

   Faculty of Health, Medicine and Life Sciences, School for Mental Health and Neuroscience, Alzheimer Centre Limburg, Maastricht University, Maastricht, Netherlands

   Contribution

   Data curation, Formal analysis, Investigation, Visualization, Methodology, Data interpretation

   Contributed equally with

   Heidi IL Jacobs

   Competing interests

   No competing interests declared

3. Benedikt A Poser

   Faculty of Psychology and Neuroscience, Department of Cognitive Neuroscience, Maastricht University, Maastricht, Netherlands

   Contribution

   Software, Supervision, Data interpretation

   Competing interests

   No competing interests declared

4. Linda HG Pagen

   Faculty of Health, Medicine and Life Sciences, School for Mental Health and Neuroscience, Alzheimer Centre Limburg, Maastricht University, Maastricht, Netherlands

   Contribution

   Investigation

   Competing interests

   No competing interests declared

5. Dimo Ivanov

   Faculty of Psychology and Neuroscience, Department of Cognitive Neuroscience, Maastricht University, Maastricht, Netherlands

   Contribution

   Software, Data interpretation

   Competing interests

   No competing interests declared

6. Frans RJ Verhey

   Faculty of Health, Medicine and Life Sciences, School for Mental Health and Neuroscience, Alzheimer Centre Limburg, Maastricht University, Maastricht, Netherlands

   Contribution

   Data interpretation

   Competing interests

   No competing interests declared

7. Kâmil Uludağ

   1. Center for Neuroscience Imaging Research, Institute for Basic Science and Department of Biomedical Engineering, Sungkyunkwan University, Suwon, Republic of Korea
   2. Techna Institute & Koerner Scientist in MR Imaging, University Health Network, Toronto, Canada

   Contribution

   Data interpretation

   Competing interests

   No competing interests declared

• 4,107

  views

• 443

  downloads

• 46

  citations

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