Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=pcns20

Cognitive NeuroscienceCurrent Debates, Research & Reports

ISSN: 1758-8928 (Print) 1758-8936 (Online) Journal homepage: http://www.tandfonline.com/loi/pcns20

Entrainment enhances theta oscillations andimproves episodic memory

Brooke M. Roberts, Alex Clarke, Richard J. Addante & Charan Ranganath

To cite this article: Brooke M. Roberts, Alex Clarke, Richard J. Addante & Charan Ranganath(2018) Entrainment enhances theta oscillations and improves episodic memory, CognitiveNeuroscience, 9:3-4, 181-193, DOI: 10.1080/17588928.2018.1521386

To link to this article: https://doi.org/10.1080/17588928.2018.1521386

Accepted author version posted online: 10Sep 2018.Published online: 10 Sep 2018.

Submit your article to this journal

Article views: 135

View Crossmark data

ARTICLE

Entrainment enhances theta oscillations and improves episodic memoryBrooke M. Robertsa, Alex Clarke b,c, Richard J. Addanted and Charan Ranganatha,e

aDepartment of Psychology, University of California at Davis, Davis, CA, USA; bDepartment of Psychology, University of Cambridge,Cambridge, UK; cDepartment of Psychology, Anglia Ruskin University, Cambridge, UK; dDepartment of Psychology, California StateUniversity, San Bernardino, CA, USA; eCenter for Neuroscience, University of California at Davis, Davis, CA, USA

ABSTRACTNeural oscillations in the theta band have been linked to episodic memory, but it is unclearwhether activity patterns that give rise to theta play a causal role in episodic retrieval. Here, weused rhythmic auditory and visual stimulation to entrain neural oscillations to assess whethertheta activity contributes to successful memory retrieval. In two separate experiments, humansubjects studied words and were subsequently tested on memory for the words (‘item recogni-tion’) and the context in which each had been previously studied (‘source memory’). Betweenstudy and test, subjects in the entrainment groups were exposed to audiovisual stimuli designedto enhance activity at 5.5 Hz, whereas subjects in the control groups were exposed to white noise(Expt. 1) or 14 Hz entrainment (Expt. 2). Theta entrainment selectively increased source memoryperformance in both studies. Electroencephalography (EEG) data in Expt. 2 revealed that thetaentrainment resulted in band-specific enhancement of theta power during the entrainmentperiod and during post-entrainment memory retrieval. These results demonstrate a direct linkbetween theta activity and episodic memory retrieval. Targeted manipulation of theta activitycould be a promising new approach to enhance theta activity and memory performance inhealthy individuals and in patients with memory disorders.

ARTICLE HISTORYRecieved 22 August 2018Published online 24Septemper 2018

KEYWORDSTheta; episodic memory;source memory;entrainment; EEG

Introduction

In the rat hippocampus, theta oscillations temporallystructure the activity of neural ensembles during mem-ory retrieval and spatial navigation (Hasselmo, Hay, Ilyn,& Gorchetchnikov, 2002), and these oscillations havebeen linked to successful learning and retention(Landfield, McGaugh, & Tusa, 1972; Winson, 1978).Theta activity is also prominent in human EEG record-ings in the 4–8 Hz frequency range, and studies havelinked scalp-recorded theta to medial prefrontal (Asada,Fukuda, Tsunoda, Yamaguchi, & Tonoike, 1999; Ishiiet al., 1999) and parietal sources (Foster & Parvizi, 2012;see Hsieh & Ranganath, 2014, for review) that are knownto exhibit high functional connectivity with the hippo-campus (Kahn, Andrews-Hanna, Vincent, Snyder, &Buckner, 2008; Libby, Ekstrom, Ragland, & Ranganath,2012; Wang, Ritchey, Libby, & Ranganath, 2016; Wanget al., 2014; Wang & Voss, 2015). EEG studies show thattheta activity is enhanced during successful memoryencoding and retrieval, particularly on context-depen-dent memory tasks that are thought to depend onhippocampal function (Addante, Watrous, Yonelinas,

Ekstrom, & Ranganath, 2011; Anderson & Hanslmayr,2014; Clarke, Roberts, & Ranganath, 2018; Crivelli-Decker, Hsieh, Clarke, & Ranganath, 2018; Foster,Kaveh, Dastjerdi, Miller, & Parvizi, 2013; Fuentemilla,Barnes, Duzel, & Levine, 2014; Haenschel et al., 2009;Hsieh & Ranganath, 2014; Jacobs, Hwang, Curran, &Kahana, 2006; Klimesch, Doppelmayr, Schimke, &Ripper, 1997; Rutishauser, Ross, Mamelak, & Schuman,2010; Staudigl & Hanslmayr, 2013).

If theta activity contributes to episodic memoryretrieval, then direct manipulations of theta activitycould significantly affect memory performance. Suchevidence would suggest that theta oscillations arenot mere epiphenomena (see Yartsev, Witter, &Ulanovsky, 2011), and moreover, that targetedenhancement of theta activity could be used toimprove episodic memory in healthy individuals orin patients with memory disorders.

Here, we tested whether rhythmic presentation ofauditory and visual stimuli could be used to stimulatethe brain networks that generate theta activity,through a process known as ‘oscillatory entrainment’

CONTACT Charan Ranganath cranganath@ucdavis.edu Department of Psychology, University of California at Davis, 1544 Newton Court, Davis, CA95618, USA

COGNITIVE NEUROSCIENCE2018, VOL. 9, NOS. 3–4, 181–193https://doi.org/10.1080/17588928.2018.1521386

© 2018 Informa UK Limited, trading as Taylor & Francis Group

(Calderone, Lakatos, Butler, & Castellanos, 2014;Herrmann, Struber, Helfrich, & Engel, 2015; Lakatos,Karmos, Mehta, Ulbert, & Schroeder, 2008; Thut,Schyns, & Gross, 2011; Walter, 1953). In studies ofvisual processing, entrainment has been shown toenhance oscillatory activity and modulate spikingactivity and behavioral performance (Calderoneet al., 2014; Lakatos et al., 2008). These findings sug-gest that rhythmic auditory and visual stimuli canentrain oscillatory activity, and that processing of sti-muli during entrainment can be enhanced if stimuliare delivered at the optimal phase of the ongoingoscillation. Here, we were interested in determiningwhether entrainment can be used to stimulate brainnetworks that generate theta oscillations. Specifically,we hypothesized that entrainment of theta oscilla-tions would have lasting effects on episodic memoryretrieval and on spontaneous theta activity.

To address this question, we conducted two experi-ments in which participants were exposed to audio-visual entrainment (AVE) at theta frequency (5.5 Hz)between the study (encoding) and test (retrieval)phases of an episodic memory paradigm. In a studyusing a similar memory paradigm, we (Addante et al.,2011) found that frontal and parietal theta oscillationswere enhanced during successful retrieval of contex-tual information associated with learned items(‘source memory’), but no effects were evident duringsuccessful recognition of items in the absence of cor-rect source memory retrieval (‘item memory’). Theseresults led us to predict that, if AVE has lasting effectson networks that generate theta oscillations, then wewould expect to see improved source memory perfor-mance and enhanced theta activity during sourcememory retrieval.

Materials and methods

Participants

Experiment 1: 50 right-handed subjects (26 females)ages 18–25 (average 20.3 ± 0.31 years) were recruitedfrom the UC Davis undergraduate community. 25 sub-jects each were assigned to the theta (5.5 Hz) entrain-ment group and to the control (white noise) condition.At the point of data analysis, 5 subjects were removeddue to chance source memory discriminability. Chancediscriminability was defined as Percent Source Correct—Percent Source Incorrect ≤ 1%. The final analysis for the

theta entrainment group included 24 subjects, and 21subjects were included in the final analysis for the con-trol group. However, inclusion of these subjects wouldnot change the results–all of the major findings werestatistically significant even when the entire sample wasincluded.

Experiment 2: 40 right-handed subjects (22 females)ages 18–26 (average 20.5 ± 0.27 years) were recruitedfrom the UC Davis undergraduate community. 20 sub-jects each were assigned to the 5.5 Hz entrainmentgroup and the 14 Hz entrainment control group.

Both experiments were approved by theInstitutional Review Board at the University ofCalifornia, Davis. Written informed consent wasobtained from each subject for both experimentsbefore participation.

Task

During the study phase (i.e., ‘encoding blocks’), sub-jects incidentally learned 200 words in the context oftwo different incidental encoding tasks. Words werestudied in 4 blocks, with each block consisting of 50encoding trials. During two study blocks, subjectsindicated whether each word corresponded to a liv-ing thing (for the ‘alive’ task), and during two blocks,they indicated whether it corresponded to some-thing that is man-made (for the ‘man-made’ task).The order of ‘alive’ and ‘manmade’ blocks was coun-terbalanced across subjects. Following the studyblocks, subjects received 36 minutes of either thetaentrainment at 5.5 Hz, or control (white noise inStudy 1 and 14 Hz entrainment in Study 2).

At test (i.e. retrieval), stimuli presented during theencoding phase were randomly intermixed with anadditional 100 new words (lures), for a total of 300test stimuli. Retrieval was divided into 6 testingblocks, each consisting of 50 words. Each trial com-menced with presentation of a fixation cross in thecenter of the screen for 1000 ms, followed by centralpresentation of a word for 1500 ms. Next, partici-pants were prompted to make self-paced item andsource decisions (see Figure 1). For item-recognitionjudgments, subjects responded using a 5-point con-fidence scale, with 5 indicating that they were sure itwas studied, 4 indicating that it was probably stu-died, 3 indicating guessing, 2 indicating it was prob-ably new, and 1 indicating that they were sure it wasnew. For source decisions, subjects again used a 5-

182 B. M. ROBERTS ET AL.

point confidence scale, with 5 indicating that theywere sure the word was presented in the man-madetask, 4 indicating it was probably from the man-made task, 3 indicating guessing, 2 indicating itwas probably from the alive task, and 1 indicatingthat they were sure it was from the alive task. Duringsource judgments subjects were instructed torespond with 3 for new items. Studied items thatreceived 5 or 4 responses for item judgments weretreated as item recognition ‘hits’. Source memory hitscorresponded to items studied in the alive task thatreceived 1 or 2 responses and items studied in theman-made task that received 4 or 5 responses.Percentages were calculated based on the total num-ber of trials, excluding missed trials in which subjectsfailed to make a response, and trials in which sub-jects made a 3 response (guesses). Item Onlyresponses included trials with correct item hits thatwere not associated with correct source hits.

AVE procedure

Between encoding and retrieval phases subjectsreceived 36 minutes of either theta entrainment at

5.5 Hz, or control (white noise in Study 1 and 14 Hzentrainment in Study 2). Subjects were presentedwith lights and sounds targeting a frequency of5.5 Hz, or control, in order to drive theta oscillationsprior to memory retrieval. A commercially availabledevice (DAVID PAL 36 by Mind Alive, Inc.,Edmonton, Canada) was used to administer visualstimulation (via LED lights embedded on the inneredge of darkened eyeglasses) and isochronic pulsedtones (via headphones) targeting 5.5Hz (thetaentrainment groups) or 14 Hz (control group forExperiment 2) for 36 minutes. The protocols, andduration, are pre-set in the device which deliverssynchronous, in phase, audio-visual stimulation. Forthe control condition in Experiment 1, subjects lis-tened to white noise for 36 minutes using the sameheadphones. Subjects also wore the eyeglasses withthe lights turned off. All subjects were instructed toattend to the lights and sounds, and to relax duringentrainment, but not to fall asleep. Subjects couldclose their eyes or leave them slightly open toensure the comfort of each participant during theentrainment period, as the lights remain visible witheyes closed.

Figure 1. Schematic of the episodic memory task. Participants incidentally learned lists of visually-presented words while perform-ing two semantic decision tasks. During ‘alive’ blocks, they indicated whether each word represented a living thing, and during‘man-made’ blocks, they indicated whether it represented something that was man-made. Following 36 minutes of thetaentrainment or a control condition, subjects completed a surprise memory test that requireditem and source recognition confidencejudgments. In Expt. 2, EEG was recorded during entrainment and during memory retrieval.

COGNITIVE NEUROSCIENCE 183

EEG recordings

In Experiment 2, EEG was recorded at a sampling rateof 1024 Hz with respect to a common mode sense(CMS) active electrode located on the scalp nearelectrode site Cz during the entrainment periodand subsequent memory test using a BioSemi(http://www.biosemi.com) Active Two system. EEGrecordings were taken from 64 active Ag/AgCl scalpelectrodes embedded in an elastic cap, with elec-trode locations corresponding to an extended ver-sion of the international 10/20 system (Klem, Luders,Jasper, & Elger, 1999). Recordings were also takenfrom active electrodes placed on the left and rightmastoids, and horizontal and vertical electroculo-gram (EOG) recordings were made from electrodesplaced lateral to each eye, and above and below theleft eye. The skin under the electrodes was slightlyabraded using a blunt applicator needle, and electro-des were filled with electrolyte gel. Testing was con-ducted in a sound-attenuated testing chamber.Subjects were instructed to minimize muscle tension,eye movements and blinking during, particularly dur-ing the fixation periods and test stimulus presenta-tion. Following the response period for each trial aslide was inserted on the screen indicating that itwas ‘OK to Blink.’ Participants were encouragedrestrict any blinks or eye movements to this period.This time period was not included in EEG analysis.

EEG data analysis

Power estimates during entrainment

The EEGLAB toolbox (Delorme & Makeig, 2004) inMatlab was used in EEG data analysis. Examinationof the spectral characteristics of the AVE signalsrevealed that, stimulation frequencies fluctuated inthe first 10–15 minutes, whereas stimulation wasfocused on the targeted frequency during the15–30 minutes following the onset of AVE.Accordingly, analyses of EEG data during entrain-ment focused on this epoch, when AVE was specifi-cally focused on the targeted frequency.

Our analyses of EEG activity during AVE focusedon effects on oscillatory power. We were primarilyinterested in testing the hypothesis that AVE stimu-lation would enhance EEG power at the targetedentrainment frequency. The AVE device used in thisstudy was not equipped to generate precise records

of stimulus onsets that could be synchronized to theEEG recordings, so we could not determine thephase relationship between the AVE stimuli and theevoked EEG signal. Although the phase relationshipbetween AVE and EEG activity might be relevant forunderstanding effects of entrainment on online pro-cessing, in the present study, we were interested inthe effects of AVE on behavior and activity duringthe post-entrainment period. Accordingly, the phaseof EEG evoked during AVE was not relevant to ourexperimental questions.

EEG data were resampled to 1000 Hz and filteredbetween 1 and 55 Hz using a Blackman windowedsinc FIR filter (transition bandwidth 0.2 Hz, length27,500 points). Data were segmented into 2 secondconsecutive, non-overlapping epochs and eachepoch was de-meaned. Bad channels were detectedby manual inspection of the data, and were removedbefore data were re-referenced to the average of theleft and right mastoids. Independent componentanalysis (ICA) was run using runica (Delorme,Sejnowski, & Makeig, 2007; Jung et al., 2000a,b),and SASICA (Chaumon, Bishop, & Busch, 2015) wasused for the detection of artifactual components toreject, which were visually inspected for validity. Thepower spectra were calculated using Bartlett’smethod which averages together the non-overlap-ping periodograms given by each epoch.Specifically, spectral power was calculated for eachepoch using a fast Fourier transform (FFT) for fre-quencies between 1 and 20 Hz, in 0.5 Hz steps,normalized using a log transform and averagedover epochs.

Time-frequency analysis during memory retrieval

Data analyses were performed offline using theEEGLab Toolbox (Delorme & Makeig, 2004) andFieldtrip Toolbox (Oostenveld, Fries, Maris, &Schoffelen, 2011). Analysis during memory retrievalfocused on trials with correct item recognition only,trials with correct item recognition and source mem-ory performance, and correct rejections. All othertrials were excluded from analysis. EEG data weredown-sampled to 512 Hz, re-referenced to the aver-age of the left and right mastoids, and high-passfiltered at 0.5 Hz. Data were segmented into 3.65sepochs following starting 1.25s prior to the onset ofthe word, and extending to 2.4s following the

184 B. M. ROBERTS ET AL.

presentation of the word for each trial in order toencompass the baseline period, in which subjectsfocused on a fixation cross in the center of the screenthrough the test stimulus presentation and responseperiods. These epochs were visually inspected toreject any trials that appeared to contain excessivenoise. Independent component analysis (ICA) wasused to correct for artifacts in the remaining trials(Delorme et al., 2007; Jung et al., 2000a). Epochswere visually inspected again following ICA to rejectany remaining trails that appeared to contain resi-dual EOG, EMG, skin potential, or other artifacts.Following artifact rejection, the remaining trialswere included in time-frequency analysis.

Time-frequency analysis was performed using 5.7cycle Morlet wavelet decomposition (Roach &Mathalon, 2008) ranging from 2 to 50 Hz at 47 loga-rithmically spaced frequencies. The decompositionwas performed on epochs for each trial included inanalysis. Time-frequency analysis was performed from450 ms prior to the test stimulus presentation to1600 ms following the presentation of the test stimu-lus. Data were baseline corrected to the fixation per-iod prior to test stimulus presentation (−450 to 0 ms).

After Morlet wavelet decomposition of each trial,oscillatory power was averaged for correct itemrecognition only trials and correct item recognitionand source memory trials for each subject in boththe 5.5 Hz and 14 Hz entrainment groups, resultingin condition specific time-frequency representations.

Statistical analysis

Statistical analyses were performed with Fieldtrip(Oostenveld et al., 2011) on frequencies of interest.For the analysis of the entrainment period, the datawere analyzed at 5.5 Hz and at 14 Hz. For the retrie-val data, theta activity (4–6 Hz) and low beta activity(12–16 Hz) were analyzed. Differences between theentrainment and control groups were tested using acluster-based permutation approach to identify sig-nificant electrode clusters within the frequency band(Maris & Oostenveld, 2007). Differences between con-ditions were established using a two-tailed indepen-dent t-test with an alpha of 0.05, and the sum ofcluster-forming t-values were calculated based onneighboring electrodes (with a minimum of 2 signifi-cant electrodes forming a connection). Group assign-ments were randomly shuffled and the maximum

cluster mass was retained from the permuted data.The data were permuted 5000 times to create apermutation distribution that was used to calculatesignificance values for observed clusters in the origi-nal data.

For the EEG data during entrainment, we testedfor topographic effects, and for the retrieval EEG datawe tested for topographic effects over time, between0.25 seconds and 1 second. The time period from 0to 0.25 seconds was not analyzed due to possibleinfluences of pre-stimulus activity on wavelet powerestimates.

Results

Overview

Two experiments were conducted to assess effects ofentrainment (see Figure 1). Experiment 1 was a beha-vioral study, in which item and source recognition wascompared between participants exposed to 5.5 Hzentrainment and participants who were exposed towhite noise. Experiment 2 used the same design,except that participants undergoing 5.5 Hz entrain-ment were compared against a 14 Hz entrainmentcontrol group, and EEG data were collected duringentrainment and recognition.

Experiment 1

As noted earlier, theta activity is enhanced duringsource memory retrieval tasks that are thought todepend on the hippocampus. We predicted that AVEwould stimulate the cortical networks that generatetheta, thereby increasing the likelihood of thetastates and enhancing source memory performanceduring the retrieval test. Consistent with this predic-tion, results from Experiment 1 (Figure 2(a) andTable 1) showed that the proportion of trials asso-ciated with correct Item and Source memory judg-ments (‘Item + Source’) was significantly higher forthe entrainment group compared to the control(white noise) group (F [1,43] = 4.30; p = 0.044;Cohen’s d = 0.628). There were no significant differ-ences between the entrainment and control groupsin the proportion of trials for correct ‘Item Only’ trials(F [1,43] = 1.26; p = 0.725). The rate of false alarms foritem recognition judgments also did not significantlyvary between theta entrainment and control groups

COGNITIVE NEUROSCIENCE 185

(F [1,43] = 0.474; p = 0.495). Consistent with thesefindings, signal-detection-based analyses revealedthat no significant between-group differences initem recognition discriminability (i.e. d’; F[1,43] = 2.22; p = 0.14), or response bias (F[1,43] = 0.46; p = 0.50). These results indicate thattheta entrainment increased the likelihood of suc-cessful source memory retrieval, consistent with pre-vious findings demonstrating a specific relationshipbetween theta activity and source memory retrieval(Addante et al., 2011).

According to dual-process models of recognitionmemory, Item-Only responses are primarily sup-ported by familiarity, whereas Item + Sourceresponses reflect familiarity and recollection (seeMickes, Wais, & Wixted, 2009; Yonelinas, 2002). Ifone assumes that recollection and familiarity areindependent, Item-Only hit rates underestimate theeffect of familiarity, because they do not take intoaccount the fact that familiarity cannot be estimatedfor Item + Source hits (see Yonelinas & Jacoby, 2012).Adapting the equation from Yonelinas and Jacoby(2012), we estimated familiarity as follows: p(Familiarity) = p(ItemOnly)/[1-p(Item + Source)].Consistent with our analyses of Item-Only hit rates,this analysis also did not reveal significant effects of

entrainment on familiarity (Entrainment: M = 0.91;SEM = 0.023; Control: M = 0.88; SEM = 0.044: F[1,43] = 0.320; p = 0.575).

Experiment 2

In Experiment 2, we sought to: (1) replicate theeffects of theta entrainment observed in the firstexperiment, (2) examine theta activity duringentrainment and during the subsequent memorytest, and (3) eliminate the possibility that between-group differences were driven by factors related tothe white noise control condition, rather thanentrainment. The design of Experiment 2 was iden-tical to experiment 1, except that participants in thecontrol group were exposed to AVE at 14 Hz, andEEG was concurrently recorded in order to assessoscillatory activity during entrainment and duringthe subsequent retrieval phase.

Behavioral results for Experiment 2, summarized inFigure 2(b) and Table 1, replicated those of Experiment1, showing that the proportion of Item + Source trialswas significantly higher for the 5.5 Hz entrainmentgroup compared to the 14 Hz entrainment group (F[1, 38] = 5.02; p = 0.031; Cohen’s d = 0.709). Similar toexperiment 1, the proportion of correct Item Only hits

Figure 2. Theta entrainment enhances source memory performance. Proportions of Item + Source hits, Item Only hits, and FalseAlarms are plotted for (a) Experiment 1 and (b) Experiment 2. Asterisks indicate significant differences between theta entrainmentand control groups.

Table 1. Mean performance (standard error of the mean) for the entrainment and control groups for both experiments.Item + Source hits Item Only hits False Alarms Item d’ C

Experiment 1 Theta 73.8% (2.1) 11.2% (1.8) 16.7% (1.9) 1.94 (0.11) −0.24 (0.06)Control 68.3% (1.5) 12.2% (2.0) 18.8% (2.3) 1.73 (0.07) −0.31 (0.05)

Experiment 2 Theta 70.9% (1.3) 11.9% (1.7) 19.1% (3.1) 1.96 (0.14) −0.23 (0.04)Control 66.2% (1.9) 16.2% (1.9) 25.8% (2.7) 1.65 (0.10) −022 (0.04)

186 B. M. ROBERTS ET AL.

were not significantly different across the theta entrain-ment and control groups (F [1,38] = 2.93; p = 0.095).The rate of false alarms for new items also did notsignificantly differ across theta entrainment and con-trol groups (F [1,38] = 2.61; p = 0.114). Moreover, signal-detection analyses revealed no difference betweenentrainment groups in item recognition discriminabil-ity (d’; [F [1.38] = 3.28; p = 0.08] or response bias [F[1,38] = 0.01; p = 0.92]. Analyses assuming independentfamiliarity and recollection processes also revealed nosignificant effects of entrainment group on item famil-iarity estimates (Entrainment: M = 0.94; SEM = 0.027;Control: M = 0.95; SEM = 0.057; F[1,38] = 0.0006;p = 0.981). The striking parallels between the beha-vioral results of Experiments 1 and 2 demonstrate therobustness of entrainment effects on memory.Furthermore, because each study used different rigor-ous controls, findings from the two experiments sug-gest that the between-group differences were drivenby theta entrainment and not by factors peculiar to thecontrol condition.

We next analyzed the EEG data in order to deter-mine the direct effects of AVE on oscillatory activity.During entrainment, EEG power was enhanced spe-cifically at the entrainment frequencies for both the5.5 Hz entrainment group, and the 14 Hz controlentrainment group (Figure 3(a)). Statistical analysesfor power at 5.5 Hz revealed a significant enhance-ment for the 5.5 Hz entrainment group compared tothe 14 Hz control group across posterior electrodes(cluster p = 0.0064; Figure 3(b)), while power at 14 Hzwas significantly enhanced for the 14 Hz group com-pared to the 5.5 Hz group across most electrodes,

peaking over central electrodes (cluster p = 0.0036;Figure 3(c)).

Having established that entrainment directlyaffected oscillatory activity during the retentioninterval, we next tested whether entrainmentaffected endogenous EEG activity during the subse-quent memory test. If entrainment led to persistentchanges in the excitability of the networks thatgenerate theta activity, then there are at least twoways in which it could affect performance (Mitchell& Johnson, 2009; Rugg & Wilding, 2000; Wilding &Ranganath, 2011). One possibility is that thetaentrainment enhances recollective processes thatare engaged specifically following recognition (i.e.,‘Post-retrieval Monitoring’). In this case, we wouldexpect an interaction between theta entrainmentand source recognition, such that theta power inthe 5.5 Hz group would be elevated specifically onItem + Source trials (cf. Rugg & Wilding, 2000).Alternatively, 5.5 Hz entrainment could increasethe engagement of memory monitoring and controlprocesses (i.e., ‘retrieval orientation’) that areengaged prior to successful memory retrieval(Johnson, Kounios, & Nolde, 1997; Johnson &Rugg, 2006; Ranganath & Paller, 1999, 2000), inwhich case we would expect to see a main effectof entrainment group, such that theta power duringretrieval would be generally enhanced following5.5 Hz entrainment.

We first compared theta activity between the5.5 Hz entrainment group and the 14 Hz controlgroup for all trial types. Statistical analysis showedsustained theta activity increases over frontal

Figure 3. Entrainment enhances EEG power at the stimulation frequencies. (a) Power spectrum (averaged across all electrodes)showing peaks for the 5.5 Hz group (in red) and 14 Hz group (in blue) at the specific frequencies used during AVE, and effects at theharmonics. (b) Topographic map showing enhanced power at 5.5 Hz for the 5.5 Hz entrainment group compared to the 14 Hzentrainment group during AVE. Asterisks indicate the electrodes showing a significant power enhancement for the 5.5 Hz groupcompared to the 14 Hz group. (c) Topographic map showing enhanced power at 14 Hz for the 14 Hz control group compared to the5.5 Hz group. Asterisks indicate the electrodes showing a significant power enhancement for the 14 Hz group compared to the5.5 Hz group.

COGNITIVE NEUROSCIENCE 187

electrodes for the 5.5 Hz entrainment group com-pared to the 14 Hz control group (0.25–1 sec, clusterp = 0.0392; Figure 4(a)). Collapsing over time, thetaincreases were most pronounced over frontal andparietal electrodes (Figure 4(b)). These effects wereapparent only below 7 Hz, and present in both totaltheta activity (Figure 5(a); Supp. Figure 1) andinduced theta activity (Figure 5(b); Supp. Figure 2).Although beta activity appeared to be numericallyincreased at some parietal electrodes (Figure 4(c)),there were no significant effects in the beta band.We next tested for selective increases for Item +Source trials compared to Item-Only trials. Althoughnumerically larger, there were no significant differ-ences in theta activity between the 5.5 Hz group andthe 14 Hz group. Accordingly, the data are mostconsistent with an account by which theta entrain-ment modulated one’s retrieval orientation in a man-ner that enhanced successful source monitoring(Mitchell & Johnson, 2009; Rugg & Wilding, 2000).

Discussion

To summarize, results from two independent experi-ments showed that entrainment of theta oscillationsenhanced source memory performance, when com-pared against white noise (Expt. 1) or entrainment ata non-theta frequency (Expt. 2). Furthermore,entrainment at theta frequency increased sponta-neous theta activity during memory retrieval, sug-gesting that it had lasting effects on the cortico-hippocampal networks that generate theta activity.The results demonstrate that direct manipulations oftheta oscillations have relatively selective effects onmemory retrieval and endogenous neural activity.These studies indicate that theta oscillations reflectneural processes that directly contribute to success-ful episodic memory retrieval.

Previous attempts to influence neural oscillatoryactivity with entrainment have largely focused oneither short periods of flickering visual stimulation

Figure 4. Theta entrainment enhances spontaneous theta power during memory retrieval. (a) Topographies showing enhancedtheta activity for the 5.5 Hz entrainment group compared to the 14 Hz control group. Asterisks show electrodes in significant clusterover frontal electrodes. (b) Enhanced theta activity averaged across time for the 5.5 Hz entrainment group compared to the 14 Hzcontrol group. (c) Low beta activity shows no significant differences between the two groups.

Figure 5. Time frequency representations showing differences between the 5.5 Hz entrainment group and the 14 Hz control groupover frontal electrodes. Specific increased power is seen below ~ 7 Hz in both an analysis of, (a) total power, and (b) induced power.

188 B. M. ROBERTS ET AL.

(Sato, 2013; Williams, 2001; Wimber, Maass, Staudigl,Richardson-Klavehn, & Hanslmayr, 2012) or auditorytones (Henry, Herrmann, & Obleser, 2014; Henry &Obleser, 2012; Herrmann, Henry, Grigutsch, &Obleser, 2013a; Ronconi, Pincham, Cristoforetti,Facoetti, & Szucs, 2016; Will & Berg, 2007) on atrial-by-trial basis during encoding. For instance,Williams (2001) reported that flicker stimulation at10 Hz (alpha frequency) enhanced memory encod-ing. Wimber et al. (2012) used flickering visual sti-muli presented at 6 or 10 Hz to entrain oscillationsduring encoding of words. This study did not reportsignificant effects of entrainment frequency onmemory performance, but EEG power at the flickerfrequency was enhanced during memory retrieval.Another study found that flicker at 7.08 Hz inducedtransient changes in theta activity during spatialworking memory encoding, but no cognitive effectswere found for working memory performance (Sato,2013).

In the present study, extended AVE was pre-sented during the interval between memoryencoding and retrieval in a between-subjectsdesign. We chose this design because we expectedthat AVE during the retention interval would havelasting effects by enhancing the likelihood of thetastates during the subsequent memory retrievaltest. Although this hypothesis was speculative,the results of Experiment 2 demonstrated that5.5 Hz entrainment increased theta power duringthe subsequent memory retrieval test. The resultsprovide new evidence to demonstrate that pro-longed theta entrainment can have lasting effectson subsequent neurophysiology and memoryperformance.

As noted earlier, the spectral characteristics of theAVE signals from the entrainment device were notconstant throughout the entrainment period. Thesignals were delivered at variable frequencies duringthe first 15 minutes, and stabilized at 5.5 Hz from 15to 30 minutes into the entrainment session. It isconceivable that the variations in entrainment fre-quencies in the first 15 minutes could contribute tothe group differences in memory performance thatwe observed. However, this explanation does notaccount for the EEG data, which demonstrated aselective increase in power at 5.5 Hz for the thetaentrainment group during the post-entrainmentmemory retrieval task

Persistent effects of oscillatory entrainment havebeen demonstrated using other methods, such asnon-invasive brain stimulation (for review seeVeniero, Vossen, Gross, & Thut, 2015). Notably, directeffects of stimulation effects on evoked oscillatorypower typically do not persist long beyond the sti-mulation period (Hanslmayr, Matuschek, & Fellner,2014). Thus, it is more likely that persistent effectsof stimulation, and possibly entrainment, are drivenby other factors, such as NMDA receptor dependentplasticity (Nitsche et al., 2003, 2004) in prefrontal andparietal (Hsieh & Ranganath, 2014) circuits that natu-rally resonate at the stimulation frequency (Venieroet al., 2015; Vossen, Gross, & Thut, 2015; Zaehle, Rach,& Herrmann, 2010).

The present results help to clarify the role of thetaactivity in episodic memory. Many studies havereported that theta activity is enhanced during suc-cessful episodic memory retrieval (e.g., Addante,Ranganath, & Yonelinas, 2012; Addante et al., 2011;Backus, Schoffelen, Szebényi, Hanslmayr, & Doeller,2016; Clarke et al., 2018; Crivelli-Decker et al., 2018;Gruber, Tsivilis, Giabbiconi, & Müller, 2008; Guderian &Duzel, 2005; Herweg et al., 2016; Meyer, Grigutsch,Schmuck, Gaston, & Friederici, 2015), but the func-tional significance of these enhancements is unclear.Our EEG analyses provide some insights into the cog-nitive processes that were impacted by theta entrain-ment. The source memory task used here requiresboth recollection of past events (Yonelinas, 2002)and the accurate monitoring and evaluation of recol-lected details (Johnson, Raye, Mitchell, & Ankudowich,2012; Mitchell & Johnson, 2009). Because we did notfind a selective effect of entrainment on source mem-ory, the results do not provide evidence for the ideathat theta activity solely enhanced memory consolida-tion by stimulating reactivation of studied information(Fuentemilla, Penny, Cashdollar, Bunzeck, & Duzel,2010) or that it enhanced recollective processes suchas post-retrieval monitoring (Mitchell & Johnson, 2009;Rugg & Wilding, 2000). In the absence of such a selec-tive effect, the overall enhancement of theta power inthe 5.5 Hz entrainment group is more consistent withenhancement of processes that generally enhancesource monitoring. For instance, entrainment couldhave led participants to change their ‘retrieval orienta-tion’ (Rugg & Wilding, 2000) and selectively attend toretrieved details that are particularly diagnostic tosuccessful source discrimination (Mitchell & Johnson,

COGNITIVE NEUROSCIENCE 189

2009; Ranganath & Paller, 1999, 2000; Wilding &Ranganath, 2011). It is also possible that 5.5 Hzentrainment enhanced the ability to engage in a dif-ferent kind of preparatory state that enhances thelikelihood of recollection (‘retrieval mode’; cf. Lepage,Ghaffar, Nyberg, & Tulving, 2000). Both of these con-cepts can account for the selective effect of thetaentrainment on source memory accuracy and theoverall effect of entrainment on theta activity duringmemory retrieval.

The present results are somewhat discrepant withfindings from a recent study that investigated theeffects of 5 sessions of transcranial magnetic stimula-tion (TMS) to parietal cortex (Nilakantan, Bridge,Gagnon, VanHaerents, & Voss, 2017). Following up onwork by Wang, Voss, and colleagues (Wang et al., 2014;Wang & Voss, 2015), Nilakantan et al. targeted stimula-tion to parietal regions that show high functional con-nectivity with the hippocampus. Wang et al. (2014;Wang & Voss, 2015) previously showed that, followingthe stimulation sessions, those who received parietalstimulation showed better performance on an episo-dic memory test and increased intrinsic functionalconnectivity within a posterior medial cortico-hippo-campal network (Libby et al., 2012; Ranganath &Ritchey, 2012) that includes potential sources ofscalp-recorded theta activity in humans (Hsieh &Ranganath, 2014). Using a similar procedure, but adifferent memory paradigm, Nilakantan et al. foundthat stimulation did not increase the likelihood ofrecollection, but it increased the precision of recol-lected memories. Surprisingly, participants in the sti-mulation group showed a reduction in theta activityduring memory retrieval. Although it is surprising thattheir manipulation reduced theta activity andenhanced the precision of recollection, there aremany possible explanations for their results. One pos-sibility is that their reported enhancement in recollec-tion was driven by effects other than theta, such asenhancement of sub-delta (<1 Hz) oscillations thatcould mediate intrinsic functional connectivity mea-sured by fMRI (He et al., 2008). Another, non-exclusive,possibility is that theta activity promotes engagementof cognitive control processes. If parietal TMSenhanced hippocampal function, as suggested byWang et al. (2014; Wang & Voss, 2015), then thismight have enhanced spontaneous recollection andreduced the need for cognitive control during retrie-val. This account is totally speculative, but it can

potentially explain the reduction in theta activity fol-lowing stimulation in Nilakantan et al. (2017), alongwith the general finding that theta activity is asso-ciated with improved source monitoring (Mitchell &Johnson, 2009).

One potential implication of the present results isthat enhancement of oscillatory activity might be apromising target for improving learning and memoryperformance. As a noninvasive approach that does notrequire the administration of electrical current orstrong magnetic fields, AVE could be promising forpatients with memory disorders that are characterizedby abnormal oscillatory patterns (Voytek & Knight,2015). Further research is needed, however, to testthe generalizability of the reported effects to otheroscillatory entrainment methods and to other memoryparadigms. Also, it will be important to test whetherrepeated theta entrainment could elicit lasting effectson episodic memory and endogenous theta activity inhealthy individuals and in clinical populations.

Acknowledgments

We thank Steve Luck and Mariam Aly for helpful suggestionson previous drafts of this manuscript. This project was sup-ported by a Guggenheim Fellowship and by a Vannevar BushFellowship (Office of Naval Research Grant N00014-15-1-0033). Any opinions, findings, and conclusions or recommen-dations expressed in this material are those of the authorsand do not necessarily reflect the views of the Office of NavalResearch or the U.S. Department of Defense.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by the Vannevar Bush Fellowship[Office of Naval Research Grant N00014-15-1-0033].

ORCID

Alex Clarke http://orcid.org/0000-0001-7768-5229

References

Addante, R. J., Ranganath, C., & Yonelinas, A. P. (2012). ExaminingERP correlates of recognition memory: Evidence of accurate

190 B. M. ROBERTS ET AL.

source recognition without recollection. Neuroimage, 62(1),439–450.

Addante, R. J., Watrous, A. J., Yonelinas, A. P., Ekstrom, A. D., &Ranganath, C. (2011). Prestimulus theta activity predicts cor-rect source memory retrieval. Proceedings of the NationalAcademy of Sciences of the United States of America, 108,10702–10707. Epub 1014522011 Jun 1014528113

Anderson, M. C., & Hanslmayr, S. (2014). Neural mechanisms ofmotivated forgetting. Trends in Cognitive Sciences, 18, 279–292. Epub 2014 Apr 1018

Asada, H., Fukuda, Y., Tsunoda, S., Yamaguchi, M., & Tonoike,M. (1999). Frontal midline theta rhythms reflect alternativeactivation of prefrontal cortex and anterior cingulate cortexin humans. Neuroscience Letters, 274, 29–32.

Backus, A. R., Schoffelen, J. M., Szebényi, S., Hanslmayr, S., &Doeller, C. F. (2016). Hippocampal-prefrontal theta oscillationssupport memory integration. Current Biology, 26(4), 450–457.

Calderone, D. J., Lakatos, P., Butler, P. D., & Castellanos, F. X.(2014). Entrainment of neural oscillations as a modifiablesubstrate of attention. Trends in Cognitive Sciences, 18, 300–309. Epub 2014 Mar 1012

Chaumon, M., Bishop, D. V., & Busch, N. A. (2015). A practicalguide to the selection of independent components of theelectroencephalogram for artifact correction. Journal ofNeuroscience Methods, 250, 47–63. Epub 2015 Mar 1016.

Clarke, A., Roberts, B. M., & Ranganath, C. (2018). Neural oscilla-tions during conditional associative learning. NeuroImage,174, 485–493.

Crivelli-Decker, J., Hsieh, L. T., Clarke, A., & Ranganath, C. (2018).Theta oscillations promote temporal sequence learning.Neurobiology of Learning and Memory, 153, 92–103.

Delorme, A., & Makeig, S. (2004). EEGLAB: An open sourcetoolbox for analysis of single-trial EEG dynamics includingindependent component analysis. Journal of NeuroscienceMethods, 134, 9–21.

Delorme, A., Sejnowski, T., & Makeig, S. (2007). Enhanceddetection of artifacts in EEG data using higher-order statis-tics and independent component analysis. Neuroimage 34,1443–1449. Epub 2006 Dec 1426.

Foster, B. L., Kaveh, A., Dastjerdi, M., Miller, K. J., & Parvizi, J.(2013). Human retrosplenial cortex displays transient thetaphase locking with medial temporal cortex prior to activa-tion during autobiographical memory retrieval. The Journalof Neuroscience : the Official Journal of the Society forNeuroscience, 33, 10439–10446.

Foster, B. L., & Parvizi, J. (2012). Resting oscillations and cross-frequency coupling in the human posteromedial cortex.Neuroimage, 60, 384–391. Epub 2011 Dec 1029

Fuentemilla, L., Barnes, G. R., Duzel, E., & Levine, B. (2014).Theta oscillations orchestrate medial temporal lobe andneocortex in remembering autobiographical memories.Neuroimage, 85, 730–737. Epub 2013 Aug 1023

Fuentemilla, L., Penny, W. D., Cashdollar, N., Bunzeck, N., &Duzel, E. (2010). Theta-coupled periodic replay in workingmemory. Curr Biol, 20, 606–612. Epub 2010 Mar 1018

Gruber, T., Tsivilis, D., Giabbiconi, C. M., & Müller, M. M. (2008).Induced electroencephalogram oscillations during source

memory: Familiarity is reflected in the gamma band, recol-lection in the theta band. Journal of Cognitive Neuroscience,20(6), 1043–1053.

Guderian, S., & Düzel, E. (2005). Induced theta oscillationsmediate large-scale synchrony with mediotemporal areasduring recollection in humans. Hippocampus, 15, 901–912.

Haenschel, C., Bittner, R. A., Waltz, J., Haertling, F., Wibral, M.,Singer, W., . . . Rodriguez, E. (2009). Cortical oscillatory activityis critical for workingmemory as revealed by deficits in early-onset schizophrenia. The Journal of Neuroscience : the OfficialJournal of the Society for Neuroscience, 29, 9481–9489.

Hanslmayr, S., Matuschek, J., & Fellner, M. C. (2014).Entrainment of prefrontal beta oscillations induces anendogenous echo and impairs memory formation. CurrBiol, 24, 904–909. Epub 2014 Mar 1027

Hasselmo, M. E., Hay, J., Ilyn, M., & Gorchetchnikov, A. (2002).Neuromodulation, theta rhythm and rat spatial navigation.Neural Networks : the Official Journal of the InternationalNeural Network Society, 15, 689–707.

Henry, M. J., Herrmann, B., & Obleser, J. (2014). Entrainedneural oscillations in multiple frequency bands comodulatebehavior. Proceedings of the National Academy of Sciences ofthe United States of America, 111, 14935–14940. Epub1408742014 Sep 1408741129

Henry, M. J., & Obleser, J. (2012). Frequency modulationentrains slow neural oscillations and optimizes human lis-tening behavior. Proceedings of the National Academy ofSciences of the United States of America, 109, 20095–20100.Epub 1213392012 Nov 1213390114

Herrmann, B., Henry, M. J., Grigutsch, M., & Obleser, J. (2013a).Oscillatory phase dynamics in neural entrainment underpinillusory percepts of time. The Journal of Neuroscience : theOfficial Journal of the Society for Neuroscience, 33, 15799–15809.

Herrmann, C. S., Struber, D., Helfrich, R. F., & Engel, A. K. (2015). EEGoscillations: From correlation to causality. International Journalof Psychophysiology : Official Journal of the InternationalOrganization of Psychophysiology, 4, 00031–00033.

Herweg, N. A., Apitz, T., Leicht, G., Mulert, C., Fuentemilla, L., &Bunzeck, N. (2016). Theta-alpha oscillations bind the hippo-campus, prefrontal cortex, and striatum during recollection:Evidence from simultaneous EEG–FMRI. Journal ofNeuroscience, 36(12), 3579–3587.

Hsieh, L. T., & Ranganath, C. (2014). Frontal midline thetaoscillations during working memory maintenance and epi-sodic encoding and retrieval. Neuroimage, 85, 721–729.Epub 2013 Aug 1018

Ishii, R., Shinosaki, K., Ukai, S., Inouye, T., Ishihara, T., Yoshimine,T., . . . Takeda, M. (1999). Medial prefrontal cortex generatesfrontal midline theta rhythm. Neuroreport, 10, 675–679.

Jacobs, J., Hwang, G., Curran, T., & Kahana, M. J. (2006). EEGoscillations and recognition memory: Theta correlates ofmemory retrieval and decision making. Neuroimage 32,978–987. Epub 2006 Jul 2012.

Johnson, J. D., & Rugg, M. D. (2006). Electrophysiologicalcorrelates of retrieval processing: Effects of consistent ver-sus inconsistent retrieval demands. Journal of CognitiveNeuroscience, 18(9), 1531–1544.

COGNITIVE NEUROSCIENCE 191

Johnson, M. K., Kounios, J., & Nolde, S. F. (1997).Electrophysiological brain activity and memory source mon-itoring. NeuroReport, 8(5), 1317–1320.

Johnson, M. K., Raye, C. L., Mitchell, K. J., & Ankudowich, E.(2012). The cognitive neuroscience of true and false mem-ories. Nebraska Symposium on Motivation. NebraskaSymposium on Motivation, 58, 15–52.

Jung, T. P., Makeig, S., Humphries, C., Lee, T. W., McKeown, M.J., Iragui, V., & Sejnowski, T. J. (2000b). Removing electro-encephalographic artifacts by blind source separation.Psychophysiology, 37, 163–178.

Jung, T. P., Makeig, S., Westerfield, M., Townsend, J.,Courchesne, E., & Sejnowski, T. J. (2000a). Removal of eyeactivity artifacts from visual event-related potentials in nor-mal and clinical subjects. Clinical Neurophysiology : OfficialJournal of the International Federation of ClinicalNeurophysiology, 111, 1745–1758.

Kahn, I., Andrews-Hanna, J. R., Vincent, J. L., Snyder, A. Z., &Buckner, R. L. (2008). Distinct cortical anatomy linked tosubregions of the medial temporal lobe revealed by intrin-sic functional connectivity. Journal of Neurophysiology, 100,129–139. Epub 02008 Apr 00072

Klem, G. H., Luders, H. O., Jasper, H. H., & Elger, C. (1999). Theten-twenty electrode system of the international federation.The international federation of clinical Neurophysiology.Electroencephalography and Clinical Neurophysiology.Supplement, 52, 3–6.

Klimesch, W., Doppelmayr, M., Schimke, H., & Ripper, B. (1997).Theta synchronization and alpha desynchronization in amemory task. Psychophysiology, 34, 169–176.

Lakatos, P., Karmos, G., Mehta, A. D., Ulbert, I., & Schroeder, C.E. (2008). Entrainment of neuronal oscillations as a mechan-ism of attentional selection. Science, 320, 110–113.

Landfield, P. W., McGaugh, J. L., & Tusa, R. J. (1972). Thetarhythm: A temporal correlate of memory storage processesin the rat. Science, 175, 87–89.

Lepage, M., Ghaffar, O., Nyberg, L., & Tulving, E. (2000). Prefrontalcortex and episodic memory retrieval mode. Proceedings ofthe National Academy of Sciences of the United States ofAmerica, 97, 506–511.

Libby, L. A., Ekstrom, A. D., Ragland, J. D., & Ranganath, C.(2012). Differential connectivity of perirhinal and parahip-pocampal cortices within human hippocampal subregionsrevealed by high-resolution functional imaging. The Journalof Neuroscience : the Official Journal of the Society forNeuroscience, 32, 6550–6560.

Maris, E., & Oostenveld, R. (2007). Nonparametric statisticaltesting of EEG-and MEG-data. Journal of NeuroscienceMethods, 164(1), 177–190.

Meyer, L., Grigutsch, M., Schmuck, N., Gaston, P., & Friederici,A. D. (2015). Frontal–Posterior theta oscillations reflectmemory retrieval during sentence comprehension. Cortex;a Journal Devoted to the Study of the Nervous System andBehavior, 71, 205–218.

Mickes, L., Wais, P. E., & Wixted, J. T. (2009). Recollection is acontinuous process: Implications for dual-process theories ofrecognition memory. Psychological Science, 20(4), 509–515.

Mitchell, K. J., & Johnson, M. K. (2009). Source monitoring15 years later: What have we learned from fMRI about theneural mechanisms of source memory? PsychologicalBulletin, 135, 638–677.

Nilakantan, A. S., Bridge, D. J., Gagnon, E. P., VanHaerents, S.A., & Voss, J. L. (2017). Stimulation of the posterior cortical-hippocampal network enhances precision of memory recol-lection. Current Biology, 27(3), 465–470.

Nitsche, M. A., Frick, K., Henschke, U., Schlitterlau, A.,Liebetanz, D., Lang, N., & Paulus, W. (2003).Pharmacological modulation of cortical excitability shiftsinduced by transcranial direct current stimulation inhumans. The Journal of Physiology, 553(1), 293–301.

Nitsche, M. A., Grundey, J., Liebetanz, D., Lang, N., Tergau, F., &Paulus, W. (2004). Catecholaminergic consolidation ofmotor cortical neuroplasticity in humans. Cerebral Cortex,14(11), 1240–1245.

Oostenveld, R., Fries, P., Maris, E., & Schoffelen, J. M. (2011).FieldTrip: Open source software for advanced analysis ofMEG, EEG, and invasive electrophysiological data.Computational Intelligence and Neuroscience, (2011, 1.

Ranganath, C., & Paller, K. A. (1999). Frontal brain potentialsduring recognition are modulated by requirements toretrieve perceptual detail. Neuron, 22(3), 605–613.

Ranganath, C., & Paller, K. A. (2000). Neural correlates ofmemory retrieval and evaluation. Cognitive Brain Research,9(2), 209–222.

Ranganath, C, & Ritchey, M. (2012). Two cortical systems formemory-guided behaviour. Nature Reviews Neuroscience, 13(10), 713.

Roach, B. J., & Mathalon, D. H. (2008). Event-related EEG time-frequency analysis: An overview of measures and an analy-sis of early gamma band phase locking in schizophrenia.Schizophrenia Bulletin, 34, 907–926. Epub 2008 Aug 1096

Ronconi, L., Pincham, H. L., Cristoforetti, G., Facoetti, A., &Szucs, D. (2016). Shaping prestimulus neural activity withauditory rhythmic stimulation improves the temporal allo-cation of attention. Neuroreport, 27, 487–494.

Rugg, M. D., & Wilding, E. L. (2000). Retrieval processing andepisodic memory. Trends in Cognitive Sciences, 4, 108–115.

Rutishauser, U., Ross, I. B., Mamelak, A. N., & Schuman, E. M.(2010). Human memory strength is predicted by theta-fre-quency phase-locking of single neurons. Nature, 464, 903–907. Epub 02010 Mar 08824

Sato, N. (2013). Fast entrainment of human electroencephalo-gram to a theta-band photic flicker during successful mem-ory encoding. Frontiers in Human Neuroscience, 7, 208.eCollection 02013

Staudigl, T., & Hanslmayr, S. (2013). Theta oscillations at encod-ing mediate the context-dependent nature of human episo-dic memory. Curr Biol, 23, 1101–1106. Epub 2013 Jun 1106

Thut, G., Schyns, P. G., & Gross, J. (2011). Entrainment ofperceptually relevant brain oscillations by non-invasiverhythmic stimulation of the human brain. Frontiers inPsychology, 2, 170. eCollection 02011

Veniero, D., Vossen, A., Gross, J., & Thut, G. (2015). Lasting EEG/MEG aftereffects of rhythmic transcranial brain stimulation:

192 B. M. ROBERTS ET AL.

level of control over oscillatory network activity. Frontiers inCellular Neuroscience, 9, 477. eCollection 02015

Vossen, A., Gross, J., & Thut, G. (2015). Alpha power increaseafter transcranial alternating current stimulation at alphafrequency (alpha-tACS) reflects plastic changes ratherthan entrainment. Brain Stimulation, 8, 499–508. Epub2014 Dec 1020

Voytek, B., & Knight, R. T. (2015). Dynamic network commu-nication as a unifying neural basis for cognition, develop-ment, aging, and disease. Biological Psychiatry, 77, 1089–1097. Epub 2015 Apr 1028

Walter, W. G. (1953). The living brain. New York, NY: Norton.Wang, J. X., Rogers, L. M., Gross, E. Z., Ryals, A. J., Dokucu, M. E.,

Brandstatt, K. L., . . . Voss, J. L. (2014). Targeted enhance-ment of cortical-hippocampal brain networks and associa-tive memory. Science, 345, 1054–1057.

Wang, J. X., & Voss, J. L. (2015). Long-lasting enhancements ofmemory and hippocampal-cortical functional connectivityfollowing multiple-day targeted noninvasive stimulation.Hippocampus, 25, 877–883. Epub 22015 Feb 22411

Wang, S. F., Ritchey, M., Libby, L. A., & Ranganath, C. (2016).Functional connectivity based parcellation of the humanmedial temporal lobe. Neurobiology of Learning andMemory, 19, 00014–00017.

Wilding, E., & Ranganath, C. (2011). Electrophysiological corre-lates of episodic memory processes (pp. 373–396). Oxford

handbook of ERP components, Oxford University Press,New York.

Will, U., & Berg, E. (2007). Brain wave synchronization andentrainment to periodic acoustic stimuli. NeuroscienceLetters 424, 55–60. Epub 2007 Aug 2006.

Williams, J. H. (2001). Frequency-specific effects of flicker onrecognition memory. Neuroscience, 104, 283–286.

Wimber, M., Maass, A., Staudigl, T., Richardson-Klavehn, A., &Hanslmayr, S. (2012). Rapid memory reactivation revealedby oscillatory entrainment. Curr Biol, 22, 1482–1486. Epub2012 Jul 1412

Winson, J. (1978). Loss of hippocampal theta rhythm results inspatial memory deficit in the rat. Science, 201, 160–163.

Yartsev, M. M., Witter, M. P., & Ulanovsky, N. (2011). Grid cellswithout theta oscillations in the entorhinal cortex of bats.Nature, 479, 103–107.

Yonelinas, A. P. (2002). The Nature of Recollection andFamiliarity: A review of 30 Years of Research. Journal ofMemory and Language, 46, 441–517.

Yonelinas, A. P., & Jacoby, L. L. (2012). The process-dissociationapproach two decades later: Convergence, boundary con-ditions, and new directions. Memory & Cognition, 40(5),663–680.

Zaehle, T., Rach, S., & Herrmann, C. S. (2010). Transcranialalternating current stimulation enhances individual alphaactivity in human EEG. PLoS One, 5, e13766.

COGNITIVE NEUROSCIENCE 193