Neurobiology of Aging xxx (2015) 1e8
Contents lists available at ScienceDirect
Neurobiology of Aging journal homepage: www.elsevier.com/locate/neuaging
Transcutaneous vagus nerve stimulation boosts associative memory in older individuals Heidi I.L. Jacobs a, b, *, Joost M. Riphagen c, Chantalle M. Razat d, Svenja Wiese d, Alexander T. Sack b a Department of Psychiatry and Neuropsychology, School for Mental Health and Neuroscience (MHeNS), Alzheimer Centre Limburg, Maastricht University Medical Centre, Maastricht, The Netherlands b Department of Cognitive Neuroscience, Faculty of Psychology and Neuroscience, Maastricht University, Maastricht, The Netherlands c Department of Anaesthesiology, Sankt-Willibrord Spital, Emmerich, Germany d Department of Neuropsychology, Faculty of Psychology and Neuroscience, Maastricht University, Maastricht, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 October 2014 Received in revised form 20 February 2015 Accepted 23 February 2015
Direct vagus nerve stimulation (dVNS) is known to improve mood, epilepsy, and memory. Memory improvements have been observed in Alzheimer’s disease patients after long-term stimulation. The potential of transcutaneous vagus nerve stimulation (tVNS), a noninvasive alternative to dVNS, to alter memory performance remains unknown. We aimed to investigate the effect of a single-session tVNS on associative memory performance in healthy older individuals. To investigate this, we performed a singleblind sham-controlled randomized crossover pilot study in healthy older individuals (n ¼ 30, 50% female). During the stimulation or sham condition, participants performed an associative face-name memory task. tVNS enhanced the number of hits of the memory task, compared with the sham condition. This effect was specific to the experimental task. Participants reported few side effects. We conclude that tVNS is a promising neuromodulatory technique to improve associative memory performance in older individuals, even after a single session. More research is necessary to investigate its underlying neural mechanisms, the impact of varying stimulation parameters, and its applicability in patients with cognitive decline. Ó 2015 Elsevier Inc. All rights reserved.
Keywords: Vagus nerve Stimulation Memory Aging Memory modulation
1. Introduction Memory is crucial to our identity. We draw on our past experiences to guide present and future decisions. Memory problems are the most often expressed complaints by older people. Of all memory systems, episodic memory, the encoding and retrieval of events embedded in their spatiotemporal context, shows the strongest decline with aging (Ronnlund et al., 2005). In particular, the ability of forming face-name associations declines with aging (James et al., 2008). The root causes of age-related memory decline are incompletely understood; however, evidence suggests involvement of a downregulated neuromodulatory system (McIntyre et al., 2012). Nonetheless, the older brain is also capable to reorganize itself adaptively (Duffau, 2006); therefore, strategies
* Corresponding author at: School for Mental Health and Neuroscience (MHeNS), Alzheimer Center Limburg, Maastricht University Medical Centre, PO Box 616, 6200 MD Maastricht, The Netherlands. Tel.: þ31 43 388 40 90; fax: þ31 43 388 40 92. E-mail address: [email protected] (H.I.L. Jacobs). 0197-4580/$ e see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.neurobiolaging.2015.02.023
to enhance memory functions have gained popularity, also in nonclinical populations. Direct vagus nerve stimulation (dVNS), an invasive stimulation technique, has received little attention in healthy aging and agerelated cognitive decline. The idea of stimulating the vagus nerve to trigger the release of neuromodulators and modify brain activity has been pursued for more than a century (Groves and Brown, 2005). dVNS is an invasive surgical procedure in which a unidirectional wire is wrapped around the vagus nerve in the neck. This wire is connected to a subcutaneous battery implanted in the chest and sending an intermittently electrical current to the vagus nerve (George et al., 2000). The vagus nerve is the longest cranial nerve, and its fibers synapse bilaterally and topographically on neurons within the nucleus tractus solitarius, which in turn projects to the brainstem, in particular the locus coeruleus (McIntyre et al., 2012). Adrenergic activation of the vagus nerve is known to stimulate noradrenalin release from the locus coeruleus, which in turn activates several brain areas. Electrical stimulation of the vagus nerve through dVNS is believed to have specific effects on brain functioning through this pathway (McIntyre et al., 2012; Roozendaal and McGaugh, 2011).
2
H.I.L. Jacobs et al. / Neurobiology of Aging xxx (2015) 1e8
Fig. 1. Effect of transcutaneous vagus nerve stimulation on associative memory. (A) Study design. (B) Experimental paradigm. (C) Effects on associative memory performance: the number of hits (left) increased in the experimental (gray) compared with the sham (white) condition. There was no significant change in false alarm rate (right). Error bars represent 1 standard error, *p < 0.05. (For interpretation of the references to color in this Figure, the reader is referred to the web version of this article.)
Although dVNS is successful in decreasing the frequency and activity of seizures (DeGiorgio et al., 2000) and improving depressive symptoms (Marangell et al., 2002), a concomitant improvement in cognitive functions has also been observed (Clark et al., 1999). As these cognitive effects might be related to the improvement of depressive symptoms, it is important to further examine this in animals or patients with cognitive decline. Previous work has now provided evidence for memory-enhancing effects after dVNS in rats (Clark et al., 1995, 1998). Furthermore, a small longitudinal study in Alzheimer’s disease (AD) patients reported improvement or no decline in cognitive performance after 6e12 months of stimulation (Merrill et al., 2006; Sjogren et al., 2002). As this technique is invasive, only 10 patients could be included. Transcutaneous vagus nerve stimulation (tVNS), a noninvasive alternative for dVNS, has shown similar positive effects in reducing the frequency of epileptic seizures and improving depressive symptoms (Kraus et al., 2007; Stefan et al., 2012). Furthermore, tVNS has also positive effects on pain perception and tinnitus severity (Busch et al., 2013; Vanneste and De Ridder, 2012). The potential of tVNS to modulate memory performance remains unknown. The goal of this study is to investigate for the first time the effect of tVNS on memory performance, in particular face-name associations, in healthy older persons as a proof of concept. 2. Materials and methods
the Maastricht University, Faculty of Psychology and Neuroscience and the Faculty of Health, Medicine, and Life Sciences. The Ethical Review Board from the Faculty of Psychology and Neuroscience approved the study. Written informed consent was provided by all the participants. Participants were included if they met the following criteria: no evidence of cognitive deficits on neuropsychological screening, no presence of any neurological or psychiatric disease, no cardiac diseases (Kreuzer et al., 2012), no psychoactive medication use, no abuse of alcohol or drugs, Dutch as the mother tongue, and had to be able to give informed consent. Participants were invited twice at the same time of the day, separated by 7e10 days to avoid carryover effects. Each session was preceded and followed by a wide range of cognitive tests covering episodic memory (15-Word Learning Test [WLT]: learning and delayed recall) (Van der Elst et al., 2005), working memory (digit span forward and backward) (Bopp and Verhaeghen, 2005), language (verbal fluency) (Van der Elst et al., 2006c), attention (Concept Shifting Task) (Van der Elst et al., 2006a), information processing speed (Letter-Digit Substitution Test) (Van der Elst et al., 2006b), and executive functions (Stroop Color-Word Task) (Van der Elst et al., 2006d). All cognitive tests were presented in the same order for each participant, but different versions were applied to avoid practice effects. The Mini-Mental State Examination (Folstein et al., 1975) and the Hamilton Depression Rating Scale (Hamilton, 1960) were only assessed at the first meeting to check inclusion criteria.
2.1. Participants and cognitive assessment 2.2. Experimental paradigm Thirty participants (mean age, 60.57 [standard deviation {SD}, 2.54] years, 50% female), recruited from the community via local newspapers, participated in this single-blind sham-controlled randomized crossover design. All experiments were carried out at
Between the cognitive assessments, participants performed a face-name association memory task. Figure 1 provides an overview of the design. Visual stimuli of this task were presented using E-
H.I.L. Jacobs et al. / Neurobiology of Aging xxx (2015) 1e8
3
Fig. 2. Reported side effects of transcutaneous vagus nerve stimulation. The most common reported side effects were concentration problems and tiredness, which increased as time after the stimulation increased. As this is the case for both the sham and experimental conditions, this is most likely related to increased cognitive efforts, instead of the stimulation. Immediately after stimulation, one-third of the participants reported dizziness, but this appeared to be of short term. There were no serious side effects.
prime software (Psychology Software Tools Inc, Pittsburgh, PA, USA) and shown on a laptop with 17-inch screen positioned at eye height. Face stimuli were taken from the Extended Multimodal Face Database (XM2VTSDB, http://www.ee.surrey.ac.uk/CVSSP/ xm2vtsdb/) (Messer et al., 1999) and from the 2-dimensional Emotional Faces Stimuli database of the Brain Behavior Laboratory, University of Pennsylvania (Gur et al., 2001). Only faces with a neutral expression and in frontal orientation were selected and converted to a gray-scale image in the same resolution. To avoid the specific features that attracted attention, neck and clothes were manually removed, so that mainly the face, hair, and ears were visible. During the encoding phase, participants were presented with 60 neutral faces (50% female) in black-white on the middle of the screen accompanied with a name that was common in The Netherlands. Participants were asked to memorize the face-name association and to assess the gender of the person, to keep them focused to the task. Each facename association was shown for 5 seconds to provide sufficient encoding time. During the retrieval phase, participants were presented with 60 old (50% female) and 60 new (50% female) faces and had to decide whether they had seen this face during encoding. For faces that were judged as old, participants were further asked to indicate the correct name out of 4 options. There was no time limit for the retrieval phase. Participants responded to both encoding and retrieval conditions via designated buttons. During both encoding and retrieval, the responses and reaction time (milliseconds) were recorded. Encoding and retrieval phases were separated by the consolidation phase, in which participants were asked to sit relaxed, rest for 10 minutes, and think about nothing in particular. No tasks were given during the consolidation phase as the stimulation or sham procedure was still ongoing and
another task might activate other cognitive functions or other brain networks. 2.3. Stimulation Stimulation was provided via an ear clip using a circular electrode of 10 mm diameter connected as an anode to the stimulation device by a cable. For the experimental condition (tVNS), the ear clip was placed in the left external acoustic meatus on the inner side of the tragus, an anatomic area that receives sensory innervations mainly from the vagus nerve (Peuker and Filler, 2002). For the sham condition, we attached the ear clip to the left ear lobe. A second conventional electrocardiogram electrode (35 22 mm) with solid gel was placed on the right arm connected by copper cables to the device. The stimulation parameters were in line with the literature (Kraus et al., 2007; Polak et al., 2009): 8 Hz frequency, 5.0 mA electrical current, and 200 ms pulse width. Electrical stimuli were applied with a widely used transcutaneous electrical nerve stimulation device (TENSTem dental; Schwa-medico BV, Woudenberg, The Netherlands). Before the start of the experiment, we briefly applied the stimulation to the participant’s inner tragus to check if they could tolerate the sensation. These parameters were tolerated by all the participants. Participants and test instructors (SW and CMR) were not informed about the precise hypothesized outcome effects. The code was broken only after the analyses by the principal investigator (HILJ). Conditions were counterbalanced across participants. Stimulation was locked to the encoding and consolidation phase (17 minutes duration) to induce targeted plasticity (Hays et al., 2013). During stimulation, pulse rate was measured continuously by a Finger Pulse Oxymeter at the left middle finger (Saturatiemeter, Noordwijkerhout, The Netherlands). Before starting the experiment, there was a baseline measurement of 2 minutes. From 2 participants, the data from the pulse oxymeter for the sham
4
H.I.L. Jacobs et al. / Neurobiology of Aging xxx (2015) 1e8
Fig. 3. Heart rate variation during transcutaneous vagus nerve stimulation. Average heart frequency (HF) measurements over time for the sham and stimulation conditions. The coefficient of variation (COV) depicts the variability of heart rate.
condition were corrupted by technical problems. Immediately after the stimulation (after consolidation) and 90 minutes later, participants were asked to rate (0 ¼ no, 1 ¼ light, 2 ¼ moderate, and 3 ¼ strong) potential side effects, including headache, neck pain, dizziness, tiredness, nausea, tingling sensations, skin irritations at the ear, concentration, and mood changes (Fig. 2). 2.4. Data analyses Behavioral data analyses were performed with the Statistical Package for the Social Sciences (SPSS; IBM, New York, NY, USA), version 20.0 for Mac. Group characteristics were presented with the mean and standard deviations. Hits and false alarms (retrieval) were calculated, as they were the primary outcome measures of interest. Comparison between the sham and experimental conditions for the hits and false alarms were analyzed using the repeated-measures analyses of variance (ANOVAs) with condition as the within factor with 2 levels (sham or experimental stimulation) and hits and false alarms as measurements. Additionally, we applied a signal-detection analysis to investigate stimulationrelated differences in “old” versus “new” judgments. This analysis determines the sensitivity d0 and the response criterion c, representing the ability to classify old items as old (Green and Swets, 1966). This distinction cannot be assessed by the comparison of response accuracy. The d0 index computes the distance between the signal (old classified as old) and noise (old classified as new) distribution means in standard deviation units. Criterion c reflects the strategy of response of the participant, in which a value of 0 is the ideal respondent. Positive c values indicate a conservative response criterion (participants are less likely to report old items regardless of their actual presence), and negative c values reflect a liberal criterion. The parameters were calculated as described in Stanislaw and Todorov (1999): d0 ¼ F1 (H0 ) F1(F0 ) and c ¼ 0.5 (F1 [H0 ] þ F1 [F0 ]), where H0 is the corrected hit rate, F0 the corrected false alarm rate, and F1 is the function converting probabilities into z
scores. To protect against ceiling effects with H of 1 and F of 0 (corresponding z values would be þN or N, respectively), we used corrected values of H and F: H0 ¼ (h þ 0.5)/(h þ m þ 1) and F0 ¼ (f þ 0.5)/(f þ cr þ 1), where h is the number of hits (old as old), m the number of misses (old as new), f the number of false alarms (new as old), and cr is the number of correct rejections (new as new). Both the sensitivity index d0 and the response criterion c can be used together to compute the likelihood ratio b, another measure of the response bias: exp (d0 c). The neuropsychological performances were also analyzed with the repeated-measures ANOVA with condition (sham or experimental stimulation) and time (before or after intervention) as the within factors and the various tests as measurements. Main effects of time and condition and the interaction effect of “time by condition” were analyzed. This interaction effect would indicate whether the performance on the neuropsychological tests before and after the stimulation session was different for the experimental than for the sham condition. Effect sizes are expressed in etasquared measures. Significance threshold was set at p < 0.05. Side effects were expressed in percentages of the number of participants for each time point. Heart rate was expressed as the average heart rate across participants for each measurement (each second). To understand the degree of variation across participants, we calculated to coefficient of variation with the following formula for each measurement: standard deviation/mean 100. 3. Results 3.1. Demographic characteristics Table 1 provides an overview of the demographic characteristics of the participants. The mean age of the included participants was 60.57 (SD, 2.54) years, 50% was female, the mean Mini-Mental State Examination was 29.20 (SD, 0.92), and the mean on the Hamilton Depression Rating Scale was 1.57 (SD, 1.50). An overview of the
H.I.L. Jacobs et al. / Neurobiology of Aging xxx (2015) 1e8 Table 1 Characteristics of the participants (n ¼ 30) Characteristics
Mean (SD)
Age (y) Sex (% F) Educational level Mini-Mental State Examination (score) Hamilton Depression Rating Scale (score)
60.57 (2.54) 50 5.17 (1.72) 29.20 (0.92) 1.57 (1.50)
A standardized 8-point scale was used to indicate educational level (range: 1 ¼ primary school to 8 ¼ university). Key: F, female; SD, standard deviation.
performance of the participants on the various cognitive tasks, including the experimental task for each time point measured, is provided in Table 2. 3.2. Effect of stimulation on the face-name association memory task The repeated-measures ANOVA revealed an effect of condition on the number of hits (F1,29 ¼ 4.278, p ¼ 0.048, s2 ¼ 0.144), in which the experimental condition was associated with higher scores than the sham (Fig. 1C). No effect was found for the amount of false alarms (F1.29 ¼ 0.776, p ¼ 0.386, s2 ¼ 0.026). We found no effect of condition on the sensitivity index d’ prime (F1,29 ¼ 3.745, p ¼ 0.063, s2 ¼ 0.114), c criterion (F1,29 ¼ 3.240, p ¼ 0.082, s2 ¼ 0.100), or the b value (F1,29 ¼ 0.215, p ¼ 0.646, s2 ¼ 0.007). There were no significant differences between the sham and experimental conditions regarding the reaction times (all ps > 0.05). As the vagus nerve is associated with vasovagal responses and baroreceptors, partially responsible for maintaining blood pressure (Meel-van den Abeelen et al., 2013), we performed post hoc analyses in which we excluded participants taking antihypertensive drugs (n ¼ 5). Three participants were treated for hypertension with beta blockers and 2 participants with angiotensin II blocker. These post hoc analyses revealed a slightly larger effect of stimulation on hits compared with sham (F1,24 ¼ 4.758, p ¼ 0.039, s2 ¼ 0.198). Furthermore, we found a significant effect of condition on d’ prime (F1,24 ¼ 4.896, p ¼ 0.037, s2 ¼ 0.169). No significant effect of condition was found for
5
the reaction times (all ps > 0.05, see Table 2), the amount of false alarms (F1,24 ¼ 0.679, p ¼ 0.418, s2 ¼ 0.028), the c criterion (F1,24 ¼ 2.570, p ¼ 0.122, s2 ¼ 0.097), or the b value (F1,24 ¼ 0.058, p ¼ 0.811, s2 ¼ 0.002). 3.3. Comparison of neuropsychological scores over time and conditions The repeated-measures ANOVAs (for an overview, see also Table 2) showed, for the total learning score of the WLT, a significant main effect of time (F1,29 ¼ 8.544, p ¼ 0.007, s2 ¼ 0.228), no effect of condition (F1,29 ¼ 0.014, p > 0.05), and a significant interaction “time by condition” (F1,29 ¼ 5.033, p ¼ 0.033, s2 ¼ 0.148). For the delayed recall of the WLT, we found a main effect of time (F1,29 ¼ 9.919, p ¼ 0.004, s2 ¼ 0.342), no main effect of condition (F1,29 ¼ 0.003, p > 0.05), and a significant effect for the interaction “time by condition” (F1,29 ¼ 5.800, p ¼ 0.023, s2 ¼ 0.167). For both WLT indices, the results revealed that after the sham procedure the performance decreased compared with before the sham procedure. However, for the experimental stimulation, there was no significant decrease over time. Overall, the mean on the WLT scores did not differ between both conditions (see Table 2). For the interference condition of the Stroop Color-Word Task, we found a significant main effect of time (F1,29 ¼ 48.329, p < 0.001, s2 ¼ 1.67), not for condition (F1,29 ¼ 0.117, p > 0.05), and no interaction effect (F1,29 ¼ 0.020, p > 0.05). For the switching condition of the Concept Shifting Task, we found a significant effect of time (F1,29 ¼ 7.270, p ¼ 0.012, s2 ¼ 0.200), not for condition (F1,29 ¼ 0.685, p > 0.05), and no interaction effect (F1,29 ¼ 0.062, p > 0.05). For the digit span forward, we also found an effect of time (F1,29 ¼ 4.741, p ¼ 0.038, s2 ¼ 0.141) but not for condition (F1,29 ¼ 0.848, p > 0.05) and not for the interaction (F1,29 ¼ 3.114, p > 0.05). For the digit span backward, we found a significant main effect of time (F1,29 ¼ 5.362, p ¼ 0.028, s2 ¼ 0.165) but not for condition (F1,29 ¼ 0.967, p > 0.05) and not for the interaction “time by condition” (F1,29 ¼ 0.727, p > 0.05). For the Letter-Digit Substitution Test performance, we found no effect of time (F1,29 ¼ 1.835, p > 0.05), condition (F1,29 ¼ 0.473, p > 0.05), and no interaction effect (F1,29 ¼ 0.395, p 0.05).
Table 2 Performance on neuropsychological and experimental memory tests across conditions for the total sample (n ¼ 30) Time
Pre-sham, mean (SD)
Post-sham, mean (SD)
Prestimulation, mean (SD)
Poststimulation, mean (SD)
F test
p
F test
p
Word learning total score (number of words) Word learning delayed recall (number of words) SCWT interference (seconds card 3) LDST (correct number of items) CST interference (s) Digit span forward Digit span backward
51.23 (9.53)
45.73 (11.02)
48.30 (11.44)
48.90 (9.96)
8.544
0.007**
0.014
0.908
5.033
0.033*
10.77 (3.12)
8.63 (3.84)
9.83 (3.38)
9.53 (3.04)
9.919
0.004**
0.003
0.957
5.800
0.023*
92.95 (24.31)
85.99 (23.52)
93.81 (24.51)
86.57 (24.73)
48.329
Contents lists available at ScienceDirect
Neurobiology of Aging journal homepage: www.elsevier.com/locate/neuaging
Transcutaneous vagus nerve stimulation boosts associative memory in older individuals Heidi I.L. Jacobs a, b, *, Joost M. Riphagen c, Chantalle M. Razat d, Svenja Wiese d, Alexander T. Sack b a Department of Psychiatry and Neuropsychology, School for Mental Health and Neuroscience (MHeNS), Alzheimer Centre Limburg, Maastricht University Medical Centre, Maastricht, The Netherlands b Department of Cognitive Neuroscience, Faculty of Psychology and Neuroscience, Maastricht University, Maastricht, The Netherlands c Department of Anaesthesiology, Sankt-Willibrord Spital, Emmerich, Germany d Department of Neuropsychology, Faculty of Psychology and Neuroscience, Maastricht University, Maastricht, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 October 2014 Received in revised form 20 February 2015 Accepted 23 February 2015
Direct vagus nerve stimulation (dVNS) is known to improve mood, epilepsy, and memory. Memory improvements have been observed in Alzheimer’s disease patients after long-term stimulation. The potential of transcutaneous vagus nerve stimulation (tVNS), a noninvasive alternative to dVNS, to alter memory performance remains unknown. We aimed to investigate the effect of a single-session tVNS on associative memory performance in healthy older individuals. To investigate this, we performed a singleblind sham-controlled randomized crossover pilot study in healthy older individuals (n ¼ 30, 50% female). During the stimulation or sham condition, participants performed an associative face-name memory task. tVNS enhanced the number of hits of the memory task, compared with the sham condition. This effect was specific to the experimental task. Participants reported few side effects. We conclude that tVNS is a promising neuromodulatory technique to improve associative memory performance in older individuals, even after a single session. More research is necessary to investigate its underlying neural mechanisms, the impact of varying stimulation parameters, and its applicability in patients with cognitive decline. Ó 2015 Elsevier Inc. All rights reserved.
Keywords: Vagus nerve Stimulation Memory Aging Memory modulation
1. Introduction Memory is crucial to our identity. We draw on our past experiences to guide present and future decisions. Memory problems are the most often expressed complaints by older people. Of all memory systems, episodic memory, the encoding and retrieval of events embedded in their spatiotemporal context, shows the strongest decline with aging (Ronnlund et al., 2005). In particular, the ability of forming face-name associations declines with aging (James et al., 2008). The root causes of age-related memory decline are incompletely understood; however, evidence suggests involvement of a downregulated neuromodulatory system (McIntyre et al., 2012). Nonetheless, the older brain is also capable to reorganize itself adaptively (Duffau, 2006); therefore, strategies
* Corresponding author at: School for Mental Health and Neuroscience (MHeNS), Alzheimer Center Limburg, Maastricht University Medical Centre, PO Box 616, 6200 MD Maastricht, The Netherlands. Tel.: þ31 43 388 40 90; fax: þ31 43 388 40 92. E-mail address: [email protected] (H.I.L. Jacobs). 0197-4580/$ e see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.neurobiolaging.2015.02.023
to enhance memory functions have gained popularity, also in nonclinical populations. Direct vagus nerve stimulation (dVNS), an invasive stimulation technique, has received little attention in healthy aging and agerelated cognitive decline. The idea of stimulating the vagus nerve to trigger the release of neuromodulators and modify brain activity has been pursued for more than a century (Groves and Brown, 2005). dVNS is an invasive surgical procedure in which a unidirectional wire is wrapped around the vagus nerve in the neck. This wire is connected to a subcutaneous battery implanted in the chest and sending an intermittently electrical current to the vagus nerve (George et al., 2000). The vagus nerve is the longest cranial nerve, and its fibers synapse bilaterally and topographically on neurons within the nucleus tractus solitarius, which in turn projects to the brainstem, in particular the locus coeruleus (McIntyre et al., 2012). Adrenergic activation of the vagus nerve is known to stimulate noradrenalin release from the locus coeruleus, which in turn activates several brain areas. Electrical stimulation of the vagus nerve through dVNS is believed to have specific effects on brain functioning through this pathway (McIntyre et al., 2012; Roozendaal and McGaugh, 2011).
2
H.I.L. Jacobs et al. / Neurobiology of Aging xxx (2015) 1e8
Fig. 1. Effect of transcutaneous vagus nerve stimulation on associative memory. (A) Study design. (B) Experimental paradigm. (C) Effects on associative memory performance: the number of hits (left) increased in the experimental (gray) compared with the sham (white) condition. There was no significant change in false alarm rate (right). Error bars represent 1 standard error, *p < 0.05. (For interpretation of the references to color in this Figure, the reader is referred to the web version of this article.)
Although dVNS is successful in decreasing the frequency and activity of seizures (DeGiorgio et al., 2000) and improving depressive symptoms (Marangell et al., 2002), a concomitant improvement in cognitive functions has also been observed (Clark et al., 1999). As these cognitive effects might be related to the improvement of depressive symptoms, it is important to further examine this in animals or patients with cognitive decline. Previous work has now provided evidence for memory-enhancing effects after dVNS in rats (Clark et al., 1995, 1998). Furthermore, a small longitudinal study in Alzheimer’s disease (AD) patients reported improvement or no decline in cognitive performance after 6e12 months of stimulation (Merrill et al., 2006; Sjogren et al., 2002). As this technique is invasive, only 10 patients could be included. Transcutaneous vagus nerve stimulation (tVNS), a noninvasive alternative for dVNS, has shown similar positive effects in reducing the frequency of epileptic seizures and improving depressive symptoms (Kraus et al., 2007; Stefan et al., 2012). Furthermore, tVNS has also positive effects on pain perception and tinnitus severity (Busch et al., 2013; Vanneste and De Ridder, 2012). The potential of tVNS to modulate memory performance remains unknown. The goal of this study is to investigate for the first time the effect of tVNS on memory performance, in particular face-name associations, in healthy older persons as a proof of concept. 2. Materials and methods
the Maastricht University, Faculty of Psychology and Neuroscience and the Faculty of Health, Medicine, and Life Sciences. The Ethical Review Board from the Faculty of Psychology and Neuroscience approved the study. Written informed consent was provided by all the participants. Participants were included if they met the following criteria: no evidence of cognitive deficits on neuropsychological screening, no presence of any neurological or psychiatric disease, no cardiac diseases (Kreuzer et al., 2012), no psychoactive medication use, no abuse of alcohol or drugs, Dutch as the mother tongue, and had to be able to give informed consent. Participants were invited twice at the same time of the day, separated by 7e10 days to avoid carryover effects. Each session was preceded and followed by a wide range of cognitive tests covering episodic memory (15-Word Learning Test [WLT]: learning and delayed recall) (Van der Elst et al., 2005), working memory (digit span forward and backward) (Bopp and Verhaeghen, 2005), language (verbal fluency) (Van der Elst et al., 2006c), attention (Concept Shifting Task) (Van der Elst et al., 2006a), information processing speed (Letter-Digit Substitution Test) (Van der Elst et al., 2006b), and executive functions (Stroop Color-Word Task) (Van der Elst et al., 2006d). All cognitive tests were presented in the same order for each participant, but different versions were applied to avoid practice effects. The Mini-Mental State Examination (Folstein et al., 1975) and the Hamilton Depression Rating Scale (Hamilton, 1960) were only assessed at the first meeting to check inclusion criteria.
2.1. Participants and cognitive assessment 2.2. Experimental paradigm Thirty participants (mean age, 60.57 [standard deviation {SD}, 2.54] years, 50% female), recruited from the community via local newspapers, participated in this single-blind sham-controlled randomized crossover design. All experiments were carried out at
Between the cognitive assessments, participants performed a face-name association memory task. Figure 1 provides an overview of the design. Visual stimuli of this task were presented using E-
H.I.L. Jacobs et al. / Neurobiology of Aging xxx (2015) 1e8
3
Fig. 2. Reported side effects of transcutaneous vagus nerve stimulation. The most common reported side effects were concentration problems and tiredness, which increased as time after the stimulation increased. As this is the case for both the sham and experimental conditions, this is most likely related to increased cognitive efforts, instead of the stimulation. Immediately after stimulation, one-third of the participants reported dizziness, but this appeared to be of short term. There were no serious side effects.
prime software (Psychology Software Tools Inc, Pittsburgh, PA, USA) and shown on a laptop with 17-inch screen positioned at eye height. Face stimuli were taken from the Extended Multimodal Face Database (XM2VTSDB, http://www.ee.surrey.ac.uk/CVSSP/ xm2vtsdb/) (Messer et al., 1999) and from the 2-dimensional Emotional Faces Stimuli database of the Brain Behavior Laboratory, University of Pennsylvania (Gur et al., 2001). Only faces with a neutral expression and in frontal orientation were selected and converted to a gray-scale image in the same resolution. To avoid the specific features that attracted attention, neck and clothes were manually removed, so that mainly the face, hair, and ears were visible. During the encoding phase, participants were presented with 60 neutral faces (50% female) in black-white on the middle of the screen accompanied with a name that was common in The Netherlands. Participants were asked to memorize the face-name association and to assess the gender of the person, to keep them focused to the task. Each facename association was shown for 5 seconds to provide sufficient encoding time. During the retrieval phase, participants were presented with 60 old (50% female) and 60 new (50% female) faces and had to decide whether they had seen this face during encoding. For faces that were judged as old, participants were further asked to indicate the correct name out of 4 options. There was no time limit for the retrieval phase. Participants responded to both encoding and retrieval conditions via designated buttons. During both encoding and retrieval, the responses and reaction time (milliseconds) were recorded. Encoding and retrieval phases were separated by the consolidation phase, in which participants were asked to sit relaxed, rest for 10 minutes, and think about nothing in particular. No tasks were given during the consolidation phase as the stimulation or sham procedure was still ongoing and
another task might activate other cognitive functions or other brain networks. 2.3. Stimulation Stimulation was provided via an ear clip using a circular electrode of 10 mm diameter connected as an anode to the stimulation device by a cable. For the experimental condition (tVNS), the ear clip was placed in the left external acoustic meatus on the inner side of the tragus, an anatomic area that receives sensory innervations mainly from the vagus nerve (Peuker and Filler, 2002). For the sham condition, we attached the ear clip to the left ear lobe. A second conventional electrocardiogram electrode (35 22 mm) with solid gel was placed on the right arm connected by copper cables to the device. The stimulation parameters were in line with the literature (Kraus et al., 2007; Polak et al., 2009): 8 Hz frequency, 5.0 mA electrical current, and 200 ms pulse width. Electrical stimuli were applied with a widely used transcutaneous electrical nerve stimulation device (TENSTem dental; Schwa-medico BV, Woudenberg, The Netherlands). Before the start of the experiment, we briefly applied the stimulation to the participant’s inner tragus to check if they could tolerate the sensation. These parameters were tolerated by all the participants. Participants and test instructors (SW and CMR) were not informed about the precise hypothesized outcome effects. The code was broken only after the analyses by the principal investigator (HILJ). Conditions were counterbalanced across participants. Stimulation was locked to the encoding and consolidation phase (17 minutes duration) to induce targeted plasticity (Hays et al., 2013). During stimulation, pulse rate was measured continuously by a Finger Pulse Oxymeter at the left middle finger (Saturatiemeter, Noordwijkerhout, The Netherlands). Before starting the experiment, there was a baseline measurement of 2 minutes. From 2 participants, the data from the pulse oxymeter for the sham
4
H.I.L. Jacobs et al. / Neurobiology of Aging xxx (2015) 1e8
Fig. 3. Heart rate variation during transcutaneous vagus nerve stimulation. Average heart frequency (HF) measurements over time for the sham and stimulation conditions. The coefficient of variation (COV) depicts the variability of heart rate.
condition were corrupted by technical problems. Immediately after the stimulation (after consolidation) and 90 minutes later, participants were asked to rate (0 ¼ no, 1 ¼ light, 2 ¼ moderate, and 3 ¼ strong) potential side effects, including headache, neck pain, dizziness, tiredness, nausea, tingling sensations, skin irritations at the ear, concentration, and mood changes (Fig. 2). 2.4. Data analyses Behavioral data analyses were performed with the Statistical Package for the Social Sciences (SPSS; IBM, New York, NY, USA), version 20.0 for Mac. Group characteristics were presented with the mean and standard deviations. Hits and false alarms (retrieval) were calculated, as they were the primary outcome measures of interest. Comparison between the sham and experimental conditions for the hits and false alarms were analyzed using the repeated-measures analyses of variance (ANOVAs) with condition as the within factor with 2 levels (sham or experimental stimulation) and hits and false alarms as measurements. Additionally, we applied a signal-detection analysis to investigate stimulationrelated differences in “old” versus “new” judgments. This analysis determines the sensitivity d0 and the response criterion c, representing the ability to classify old items as old (Green and Swets, 1966). This distinction cannot be assessed by the comparison of response accuracy. The d0 index computes the distance between the signal (old classified as old) and noise (old classified as new) distribution means in standard deviation units. Criterion c reflects the strategy of response of the participant, in which a value of 0 is the ideal respondent. Positive c values indicate a conservative response criterion (participants are less likely to report old items regardless of their actual presence), and negative c values reflect a liberal criterion. The parameters were calculated as described in Stanislaw and Todorov (1999): d0 ¼ F1 (H0 ) F1(F0 ) and c ¼ 0.5 (F1 [H0 ] þ F1 [F0 ]), where H0 is the corrected hit rate, F0 the corrected false alarm rate, and F1 is the function converting probabilities into z
scores. To protect against ceiling effects with H of 1 and F of 0 (corresponding z values would be þN or N, respectively), we used corrected values of H and F: H0 ¼ (h þ 0.5)/(h þ m þ 1) and F0 ¼ (f þ 0.5)/(f þ cr þ 1), where h is the number of hits (old as old), m the number of misses (old as new), f the number of false alarms (new as old), and cr is the number of correct rejections (new as new). Both the sensitivity index d0 and the response criterion c can be used together to compute the likelihood ratio b, another measure of the response bias: exp (d0 c). The neuropsychological performances were also analyzed with the repeated-measures ANOVA with condition (sham or experimental stimulation) and time (before or after intervention) as the within factors and the various tests as measurements. Main effects of time and condition and the interaction effect of “time by condition” were analyzed. This interaction effect would indicate whether the performance on the neuropsychological tests before and after the stimulation session was different for the experimental than for the sham condition. Effect sizes are expressed in etasquared measures. Significance threshold was set at p < 0.05. Side effects were expressed in percentages of the number of participants for each time point. Heart rate was expressed as the average heart rate across participants for each measurement (each second). To understand the degree of variation across participants, we calculated to coefficient of variation with the following formula for each measurement: standard deviation/mean 100. 3. Results 3.1. Demographic characteristics Table 1 provides an overview of the demographic characteristics of the participants. The mean age of the included participants was 60.57 (SD, 2.54) years, 50% was female, the mean Mini-Mental State Examination was 29.20 (SD, 0.92), and the mean on the Hamilton Depression Rating Scale was 1.57 (SD, 1.50). An overview of the
H.I.L. Jacobs et al. / Neurobiology of Aging xxx (2015) 1e8 Table 1 Characteristics of the participants (n ¼ 30) Characteristics
Mean (SD)
Age (y) Sex (% F) Educational level Mini-Mental State Examination (score) Hamilton Depression Rating Scale (score)
60.57 (2.54) 50 5.17 (1.72) 29.20 (0.92) 1.57 (1.50)
A standardized 8-point scale was used to indicate educational level (range: 1 ¼ primary school to 8 ¼ university). Key: F, female; SD, standard deviation.
performance of the participants on the various cognitive tasks, including the experimental task for each time point measured, is provided in Table 2. 3.2. Effect of stimulation on the face-name association memory task The repeated-measures ANOVA revealed an effect of condition on the number of hits (F1,29 ¼ 4.278, p ¼ 0.048, s2 ¼ 0.144), in which the experimental condition was associated with higher scores than the sham (Fig. 1C). No effect was found for the amount of false alarms (F1.29 ¼ 0.776, p ¼ 0.386, s2 ¼ 0.026). We found no effect of condition on the sensitivity index d’ prime (F1,29 ¼ 3.745, p ¼ 0.063, s2 ¼ 0.114), c criterion (F1,29 ¼ 3.240, p ¼ 0.082, s2 ¼ 0.100), or the b value (F1,29 ¼ 0.215, p ¼ 0.646, s2 ¼ 0.007). There were no significant differences between the sham and experimental conditions regarding the reaction times (all ps > 0.05). As the vagus nerve is associated with vasovagal responses and baroreceptors, partially responsible for maintaining blood pressure (Meel-van den Abeelen et al., 2013), we performed post hoc analyses in which we excluded participants taking antihypertensive drugs (n ¼ 5). Three participants were treated for hypertension with beta blockers and 2 participants with angiotensin II blocker. These post hoc analyses revealed a slightly larger effect of stimulation on hits compared with sham (F1,24 ¼ 4.758, p ¼ 0.039, s2 ¼ 0.198). Furthermore, we found a significant effect of condition on d’ prime (F1,24 ¼ 4.896, p ¼ 0.037, s2 ¼ 0.169). No significant effect of condition was found for
5
the reaction times (all ps > 0.05, see Table 2), the amount of false alarms (F1,24 ¼ 0.679, p ¼ 0.418, s2 ¼ 0.028), the c criterion (F1,24 ¼ 2.570, p ¼ 0.122, s2 ¼ 0.097), or the b value (F1,24 ¼ 0.058, p ¼ 0.811, s2 ¼ 0.002). 3.3. Comparison of neuropsychological scores over time and conditions The repeated-measures ANOVAs (for an overview, see also Table 2) showed, for the total learning score of the WLT, a significant main effect of time (F1,29 ¼ 8.544, p ¼ 0.007, s2 ¼ 0.228), no effect of condition (F1,29 ¼ 0.014, p > 0.05), and a significant interaction “time by condition” (F1,29 ¼ 5.033, p ¼ 0.033, s2 ¼ 0.148). For the delayed recall of the WLT, we found a main effect of time (F1,29 ¼ 9.919, p ¼ 0.004, s2 ¼ 0.342), no main effect of condition (F1,29 ¼ 0.003, p > 0.05), and a significant effect for the interaction “time by condition” (F1,29 ¼ 5.800, p ¼ 0.023, s2 ¼ 0.167). For both WLT indices, the results revealed that after the sham procedure the performance decreased compared with before the sham procedure. However, for the experimental stimulation, there was no significant decrease over time. Overall, the mean on the WLT scores did not differ between both conditions (see Table 2). For the interference condition of the Stroop Color-Word Task, we found a significant main effect of time (F1,29 ¼ 48.329, p < 0.001, s2 ¼ 1.67), not for condition (F1,29 ¼ 0.117, p > 0.05), and no interaction effect (F1,29 ¼ 0.020, p > 0.05). For the switching condition of the Concept Shifting Task, we found a significant effect of time (F1,29 ¼ 7.270, p ¼ 0.012, s2 ¼ 0.200), not for condition (F1,29 ¼ 0.685, p > 0.05), and no interaction effect (F1,29 ¼ 0.062, p > 0.05). For the digit span forward, we also found an effect of time (F1,29 ¼ 4.741, p ¼ 0.038, s2 ¼ 0.141) but not for condition (F1,29 ¼ 0.848, p > 0.05) and not for the interaction (F1,29 ¼ 3.114, p > 0.05). For the digit span backward, we found a significant main effect of time (F1,29 ¼ 5.362, p ¼ 0.028, s2 ¼ 0.165) but not for condition (F1,29 ¼ 0.967, p > 0.05) and not for the interaction “time by condition” (F1,29 ¼ 0.727, p > 0.05). For the Letter-Digit Substitution Test performance, we found no effect of time (F1,29 ¼ 1.835, p > 0.05), condition (F1,29 ¼ 0.473, p > 0.05), and no interaction effect (F1,29 ¼ 0.395, p 0.05).
Table 2 Performance on neuropsychological and experimental memory tests across conditions for the total sample (n ¼ 30) Time
Pre-sham, mean (SD)
Post-sham, mean (SD)
Prestimulation, mean (SD)
Poststimulation, mean (SD)
F test
p
F test
p
Word learning total score (number of words) Word learning delayed recall (number of words) SCWT interference (seconds card 3) LDST (correct number of items) CST interference (s) Digit span forward Digit span backward
51.23 (9.53)
45.73 (11.02)
48.30 (11.44)
48.90 (9.96)
8.544
0.007**
0.014
0.908
5.033
0.033*
10.77 (3.12)
8.63 (3.84)
9.83 (3.38)
9.53 (3.04)
9.919
0.004**
0.003
0.957
5.800
0.023*
92.95 (24.31)
85.99 (23.52)
93.81 (24.51)
86.57 (24.73)
48.329
