Oscillatory correlates of autobiographical memory

Page 1

International Journal of Psychophysiology xxx (2014) xxx–xxx

INTPSY-10896; No of Pages 11

Contents lists available at ScienceDirect

International Journal of Psychophysiology

j ourna l homepage: www.e lsev ie r .com/ locate / i jpsycho

Oscillatory correlates of autobiographical memory

Gennady G. Knyazev a,⁎, Alexander N. Savostyanov a,b,c, Andrey V. Bocharov a,b, Elena A. Dorosheva a,b,Sergey S. Tamozhnikov a, Alexander E. Saprigyn a

a Institute of Physiology and Basic Medicine, Siberian Branch of the Russian Academy of Medical Sciences, Novosibirsk, Russiab Novosibirsk StateUniversity, Novosibirsk, Russiac Tomsk State University, Tomsk, Russia

Abbreviations:ANOVA,analysisof variance;DMN,defaencephalogram; ERSP, event-related spectral perturbatifMRI, functionalmagneticresonance imaging;GLM,generacomponent analysis; LORETA, low-resolution brain electmagnitoencephalogram;PFC,prefrontalcortex;SICA, spatiaysis; sLORETA, standardized low-resolutionbrain electromporal independent component analysis⁎ Corresponding author at: Institute of Physiology a

Timakova str., 4, Novosibirsk 630117, Russia. Tel.: +7 3842 54.

E-mail address: [email protected] (G.G. Knyazev).

http://dx.doi.org/10.1016/j.ijpsycho.2014.12.0060167-8760/© 2014 Elsevier B.V. All rights reserved.

Please cite this article as: Knyazev, G.G., et al10.1016/j.ijpsycho.2014.12.006

a b s t r a c t

a r t i c l e i n f o

Article history:Received 3 October 2014Received in revised form 6 December 2014Accepted 9 December 2014Available online xxxx

Keywords:Autobiographical memoryDefault mode networkEEGAlpha oscillationsIndependent component analysisSource localization

Recollection of events from one's own life is referred to as autobiographical memory. Autobiographical memoryis an important part of our self. Neuroimaging findings link self-referential processes with the default mode net-work (DMN). Much evidence coming primarily from functional magnetic resonance imaging studies shows thatautobiographicalmemory and DMNhave a common neural base. In this study, electroencephalographic data col-lected in 47 participants during recollection of autobiographical episodeswere analyzed using temporal and spa-tial independent component analyses in combination with source localization. Autobiographical rememberingwas associated with an increase of spectral power in alpha and beta and a decrease in delta band. The increaseof alpha power, as estimated by sLORETA, was most prominent in the posterior DMN, but was also observed invisual and motor cortices, prompting an assumption that it is associated with activation of DMN and inhibitionof irrelevant sensory and motor areas. In line with data linking delta oscillations with aversive states, decreaseof delta powerwasmore pronounced in episodes associatedwith positive emotions,whereas episodes associatedwith negative emotions were accompanied by an increase of delta power. Vividness of recollection correlatedpositively with theta oscillations. These results highlight the leading role of alpha oscillations and the DMN inthe processes accompanying autobiographical remembering.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Recollection of events from one's own life is referred to as autobio-graphical memory. During this recollection, the self is projected backthrough time and spatially and temporally bounded scene is reconstruct-ed. It may be accompanied by strong emotions (Spreng et al., 2008).Brain activation during resting conditions has been called the defaultmode of brain function and a set of regions showing a consistent patternof deactivation during the initiation of task-related activitywas called thedefault mode network (DMN, Raichle et al., 2001). Driven by an internalmode of cognition, the default modemay set the stage for self-projectionor scene construction (Gusnard et al., 2001; Raichle and Gusnard, 2005),thus enabling a functional overlap between autobiographical memory

ultmodenetwork; EEG, electro-ons; FDR, False Discovery Rate;l linearmodel; ICA, independentromagnetic tomography; MEG,l independentcomponentanal-agnetic tomography;TICA, tem-

nd Basic Medicine, SB RAMS,3 334 81 94; fax: +7 383 332

., Oscillatory correlates of auto

and the default mode of brain function. Indeed, meta-analyses of func-tionalmagnetic resonance imaging (fMRI) studies using activation likeli-hood estimation demonstrate that these two domains have a commonneural base (Spreng et al., 2008). However, the earlier vision of theDMN as a functionally homogeneous system that is broadly associatedwith internally directed cognition is now challenged by recent findingsshowing the functional heterogeneity of the DMN (Laird et al., 2009;Leech et al., 2011). An fMRI study of the topography and response profileof human parietal regions inside and outside the DMN during episodicmemory retrieval showed an activation of posterior nodes of the DMN,but alsomore anterior and dorsal parietal regions thatwere anatomicallyseparate from the DMN (Sestieri et al., 2011). The results of this studysuggest that DMN parietal regions directly support memory retrieval,whereas non-DMN parietal regions are more involved in post-retrievalprocesses, such as memory-based decision-making. Robust functionaldissociation within the DMN was also noted. Whereas posterior nodesof the DMNwere significantly activated duringmemory retrieval, an an-terior DMN node in medial prefrontal cortex was strongly deactivated(Sestieri et al., 2011). This latter finding demonstrates functional hetero-geneity rather than homogeneity within theDMNduring episodicmem-ory retrieval.

Autobiographical memory, however, entails a complex set of opera-tions, which besides self-reflection include episodic memory, emotion,visual imagery, attention, executive functions, and semantic processes

biographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/

Page 2

2 G.G. Knyazev et al. / International Journal of Psychophysiology xxx (2014) xxx–xxx

(Svoboda et al., 2006). Correspondingly, in fMRI studies of autobio-graphical retrieval, activation of many cortical areas beyond the DMNhas been documented. Particularly the medial temporal-lobe, and thehippocampal complex more specifically, is often referred to as the‘hub’ of the autobiographical memory system (Moscovitch et al.,2005). Furthermore, the link between controlled retrieval processes inautobiographical memory and activations in lateral prefrontal cortex(PFC), i.e., the classical working memory (WM) areas, was detected byan early positron emission tomography study (Conway et al., 1999)andwas supported by subsequentmeta-analyses (Svoboda et al., 2006).

The overwhelming majority of this evidence is based on fMRI find-ings. Understanding brain function however is not possible without acomparison of the evidence coming from different methodological do-mains. Particularly, electro/magneto-encephalographic (EEG/MEG)data are very important because thesemethodshave excellent temporalresolution and provide evidence for oscillatory processes in the brain attime scales relevant for the dynamics of cognitive operations. EEG stud-ies have revealed robust EEG correlates of memory processes. Moststudies emphasize a critical role of hippocampal theta in memoryencoding and retrieval (e.g., Buzsaki, 2005; Düzel et al., 2010; Fosteret al., 2013; Jensen and Lisman, 2005; Kahana et al., 2001; Nyhus andCurran, 2010; Stella and Treves, 2011; Vertes, 2005), although involve-ment of other rhythms is also acknowledged. It has been suggested thattheta and alpha oscillations reflect control processes in two memorysystems, a working memory and a complex knowledge system thatcomprises semantic long-term memory, respectively (Klimesch, 1999;Klimesch et al., 2010). In a series of studies, Conway and coauthorshave shown changes in slow cortical potentials while participantsretrieved autobiographical memories and then held them in mind. Inthese studies, 14 electrodes located at the 10–20 system coordinatesand referred to the linkedmastoidswere used. Both frontal and posteri-or effects were found and the results suggest that the left frontalnegativity primarily reflects cortical activation associatedwith the oper-ation of a complex retrieval process, whereas the later temporal and oc-cipital negativity reflects activation corresponding to the formation andmaintenance of a detailed memory (Conway et al., 2001, 2003).

Most of these studies were performed in the ‘pre-DMN’ era and,apart from the Conway et al.'s work, they did not specifically address au-tobiographical memory. In recent years, a number of studies have foundrobust EEG correlates of DMN activity. These studies used either simul-taneous EEG and fMRI registration or functional correlates of DMN ac-tivity that are known from fMRI studies (see Knyazev, 2013b for areview). However, an EEG confirmation (or refutation) of the link be-tween autobiographical memory and DMN is still lacking in the litera-ture, as well as the evidence on the involvement in these processesthe traditional EEG frequency bands. With regard to the latter, it is im-portant to take into account the existing evidence on EEG correlates ofDMN activity and self-referential processes in general.

Simultaneous EEG-fMRI studies have mostly shown a positive asso-ciation of DMN activitywith alpha and beta (Brookes et al., 2011; Hlinkaet al., 2010; Jann et al., 2009, 2010; Mantini et al., 2007; Mo et al., 2013;Ros et al., 2013; Sadaghiani et al., 2010, 2012;Wuet al., 2010) and a neg-ative association with theta (Meltzer et al., 2007; Mizuhara et al., 2004;Scheeringa et al., 2008;White et al., 2012) and delta (Hlinka et al., 2010)oscillations. Gammapower shows positive correlationswithDMNBOLDsignal at rest (Mantini et al., 2007), but it has been shown that low fre-quency oscillations (b20 Hz), and not gamma activity, predominantlycontribute to inter-areal BOLD correlations and influence local process-ing bymodulating gamma activity within individual areas (Wang et al.,2012). EEG correlates of self-referential processes are most frequentlyfound in brain regions overlapping with the DMN; spontaneous self-referential mentation is mostly associated with the alpha frequencyband (Cannon and Baldwin, 2012; Knyazev, 2013a, 2013b; Knyazevet al., 2011, 2012, 2013). Summing up, existing evidence appears toconverge in showing that in resting conditions, self-related thoughtsare accompanied by an increase of spectral power in cortical regions

Please cite this article as: Knyazev, G.G., et al., Oscillatory correlates of aut10.1016/j.ijpsycho.2014.12.006

overlapping with the DMN and these changes are most consistentlyfound in the alpha frequency band. The association between alpha oscil-lations and the default mode of brain function is in line withlongstanding observations of enhanced alpha activity duringmental im-agery and internally oriented attention (e.g. Cooper et al., 2003, 2006;Klinger et al., 1973; Ray and Cole, 1985a, 1985b). It is also in line withthe proposed role of alpha oscillations in top–down processes and inhi-bition of sensory perception (Klimesch et al., 2007).

In thus study, we aimed to investigate EEG spectral power changesin participants who retrieved autobiographical memories and thenheld them in mind. We expected that this process would be accompa-nied by an increase of spectral power in the alpha frequency band andthis increase would be most prominent in cortical regions overlappingwith the DMN. Because wewere explicitly interested in the comparisonof our results with the existing fMRI framework, the study design andmethods of EEG data analysis were chosen so that theywouldmaximal-ly match the relevant fMRI studies.

2. Material and methods

2.1. Participants

The data were collected in a sample of 47 participants (20 men; agerange of 18 to 48 years, mean = 24.2, SD = 7.5). The sample consistedof healthy volunteers who received a sum equivalent to about 5% of themonthly living wage for participation. All subjects were consistentlyright-handed. Handedness was confirmed using the Annett handednessquestionnaire and the criterion for consistent right-handedness as pro-posed by Annett (1970). All applicable subject protection guidelines andregulations were followed in the conduct of the research in accordancewith the Declaration of Helsinki. All participants gave informed consentto the study. The study has been approved by the Institute of Physiologyand Fundamental Medicine ethical committee.

2.2. Procedures

Participants sat in a soundproof and dimly illuminated room. First,the spontaneous EEG was registered during about 12 min including al-ternating 2 min intervals with eyes open and eyes closed. All followinginstructions were auditory and were given via loudspeaker. The partic-ipant was asked to close his/her eyes and after few seconds, the follow-ing instruction was given: “Please try to remember as vivid as possiblean episode from your past. Please press the button when you have cho-sen the appropriate episode and started to recall it. Please press the but-ton again when the remembrance is vivid.” To press the button, allparticipants used the index finger of his/her dominant hand. After60 s. following the second button-press, the participant was asked toevaluate the strength of positive and negative emotion and vividnessof the recollection using thefive-point Likert scale. After that, the partic-ipant performed 60 trials of a simple attention task (a GO-NO-GO task,where he/she had to press a button upon presentation of high-pitchedtone (20% of trials) and refrain from doing so upon presentation oflow-pitched tone) and after a short break proceeded to the next trialof autobiographical remembering. The insertion of intermittent GO-NO-GO task between episodes of autobiographical remembering wasmade in order to switch over participants' attention from internal to ex-ternal processing, thus allowing starting the next episode afresh. In ad-dition, in fMRI research, the switching from external to internalprocessing is the usual way to study DMN-related processes(e.g., Harrison et al., 2008). The participant was asked to keep his/hereyes closed both during the memory and during the attention tasks.Each participant completed ten paired memory-attention blocks. Be-tween each attention and memory task episodes, there was a 20 s.break, which was followed by the instruction to start the next memorytrial. For the analysis, the last but 1 s before the instruction was chosen

obiographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/

Page 3

3G.G. Knyazev et al. / International Journal of Psychophysiology xxx (2014) xxx–xxx

as the baseline and the time interval starting from 2 s before the secondbutton-press and ending 10 s after it was chosen as the test interval.

2.3. EEG recording

EEG electrodes were placed on 118 head sites according to the ex-tended International 10–10 system using the Quik-Cap128 NSL(Compumedics, Neuroscan, USA) for electrode fixation. The electroocu-logram was recorded with four electrodes placed on the orbis ocularismuscle above and below the left eye in order to detect eye-blinks, andon the left and right outer canti, approximately 1 cm lateral to eithereye, for detecting lateral and horizontal eye movements. The signalswere amplified using “Neuroscan (USA)” amplifiers, with 0.1–100 Hzanalog bandpass filter and continuously digitized at 1000 Hz. The actualpositions of each electrode and the four fiduciary points (nasion, inion,and preauricular points) were recorded by means of FASTRAK digitizer(Polhemus). All recordings were performed using a fronto-centralelectrode as ground, and electronically linkedmastoid electrodes as ref-erence. Although, in general, individual reference electrodes are recom-mended over physically linked reference electrodes (Keil et al., 2014),the linked-mastoid reference is still frequently used, particularly, inEEG studies of autobiographical memory (Conway et al., 2001, 2003)and we preserved this recording scheme for the sake of consistencywith these studies. The linked-mastoid reference was used only forEEG recording; all further analyses were performed using the averagereference. Electrode impedances were at or below 5 kilo-ohms for allelectrodes used in the analysis. EEG data were artifact-corrected usingindependent component analysis (ICA) via EEGLAB toolbox (http://www.sccn.ucsd.edu/eeglab/).

2.4. EEG data analysis

Scalp EEG samples a volume-conducted, spatially degraded versionof the electrical activity, where the potential at any location is a mixtureof multiple sources (Makeig et al., 2004; Onton et al., 2006). To over-come this limitation, different source separation and source reconstruc-tion techniques have been devised. ICA is the most popular method ofsource separation. There are two variants of ICA—spatial (SICA) andtemporal (TICA). The former is most popular in fMRI research, whereasthe latter is mostly used in EEG/MEG research. Results of SICA and TICAmay not always be similar (Peterson et al., 2000), which may explainsome inconsistencies between fMRI and EEG findings. Recently someapplications of SICA to EEG data have been proposed (Knyazev, 2013c;Knyazev et al., 2011; Ramkumar et al., 2012, 2014). A simulation studyshowed that SICA and TICA were similarly effective in determiningwidely spaced sources oscillating at different frequencies; TICA per-formed somewhat better when sources were closely spaced, whereasSICA better distinguished sources oscillating at the same frequencyand with same temporal dynamics (Knyazev, 2013c). Thus, when ap-plied to EEG data, both methods have their advantages and limitations.In this study, we aimed to apply both methods to the analysis of EEGdata and to compare the results.

Source reconstruction techniques can be roughly divided into twocategories: 3D imaging (or distributed) reconstruction methods andequivalent current dipole approaches. The former consider all possiblesource locations simultaneously, allowing for large and widely spreadclusters of activity. The latter rely on a hypothesis that only a fewsources are active simultaneously and those sources are focal. This latterassumption is unrealistic in most cases, but equivalent current dipolefitting is justified in combination with TICA (Delorme and Makeig,2004). In this study, in order to cross-validate the results, we usedboth TICA in combination with equivalent current dipole fitting andsource-space-SICA. DIPFIT EEGLAB plugin (http://sccn.ucsd.edu/wiki/A08: DIPFIT) was used for fitting equivalent current dipole models,whereas distributed source reconstruction was performed using thevery popular sLORETA package (Pascual-Marqui, 2002).

Please cite this article as: Knyazev, G.G., et al., Oscillatory correlates of auto10.1016/j.ijpsycho.2014.12.006

2.4.1. TICA approachEach participant's data consisted of ten trials; each trial included one

second of baseline and 12 s of test interval (see Section 2.2). The datawere submitted for TICA,whichwas performed for eachparticipant sep-arately using EEGLAB toolbox. The default runica algorithm (with the‘extended’, 1 option) which uses the Infomax ICA (Delorme andMakeig, 2004) was applied. DIPFIT2.2 EEGLAB plugin was used withboundary element method (BEM) head model to fit equivalent currentdipole model for each component after aligning the individual channellocations to the selected head model. Next, an EEGLAB STUDY was cre-ated, which included all datasets simultaneously, and event-relatedspectral perturbations (ERSP) were calculated for all components. TheERSP (Makeig, 1993) shows mean log event-locked deviations frombaseline-mean power at each frequency. Method of ERSP calculation re-alized in EEGLAB toolbox is described in Delorme and Makeig (2004).Time–frequency representations were calculated by Morlet waveletswith the number of cycles linearly increasing with frequency beginningat 3 cycles. For each trial, ERSP was calculated using this trial's baseline,i.e., the last but one second before the instruction in this trial. ERSPmea-sures and equivalent dipole positions were included in the pre-clustering array and an appropriate number of component clusterswas extracted using the default K-means clustering. The number of clus-terswas selected so that each cluster included components from at leasttwo thirds of all participants (i.e. at least 31 participants). To reduce ob-server bias in identifying the clusters with known functional networks,spatial correlation analysis was performed. Spatial area embracing thecluster centroid ±15 mm in all directions was correlated with referen-tial spatial templates that are described in the next section.

2.4.2. SICA approachAnalysis of ERSP in the channels space showed that most prominent

effects were observed after the second button-press (see Fig. 1). There-fore, two seconds of the test interval prior to the second button-presswere not included in SICA. For each participant, current source densityestimates in five EEG frequency bands (delta—2–4 Hz, theta—4–8 Hz,alpha—8–12 Hz, beta—12–30 Hz, and gamma—30–45 Hz) were calcu-lated for each second of each trial using sLORETA software (Pascual-Marqui, 2002). sLORETA images were converted into the NeuroimagingInformatics Technology Initiative (NIFTI) format using modifiedLOR2SPM function (http://www.ihb.spb.ru/~pet_lab/L2S/L2SMain.htm). To reveal the general effect of recollection, eleven images werecreated for each trial (i.e., each memory episode) of each participant.The first imagewas the baseline and the other images were consecutiveone-second intervals of the test period after the second button-press. Toreveal the effect of self-reported characteristics of each episode, thelog(test)—log(baseline) current source density estimateswere calculat-ed for each voxel and subsequently ten images instead of eleven werecreated for each trial. Next, group SICA was performed using theGroup spatial ICA for fMRI Toolbox (GIFT, Version 1.3i; http://icatb.sourceforge.net/), using methods and algorithms described elsewhere(Calhoun et al., 2001, 2004). Briefly, a single group ICA is performed atthe group level after subject-wise data concatenations (eachparticipant's data included all images from all 10 trials, i.e., 110 or 100images). Then obtained ICs are back reconstructed to produce single-subject time courses and spatial maps from the raw data matrix(Calhoun et al., 2001). The number of extracted componentswas initial-ly estimated from the data using the minimum description lengthscriteria (Li et al., 2007) and was subsequently set at 20.

Spatialmaps and time courses were generated for each independentcomponent and were used for component sorting. Sorting is a way toclassify the components. The components can be sorted either tempo-rally or spatially. Temporal sorting is a way to compare the model'stime course with the component's time course, i.e., to determine thedegree of ‘task-relatedness’ of each component. The components' timecourses were correlated with idealized reference variables (seebelow). Spatial sorting aims to reduce observer bias in identifying the

biographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/

Page 4

Fig. 1.Channel-level analysis. Average over channels ERSP values. For each trial, ERSPwas calculated using this trial's baseline, i.e., the last but one secondbefore the instruction in this trial.One-sample T-testwith subsequent FDR correctionwas applied to reveal significant effects. All non-significant values are zeroed out (i.e., shown in green).Warm colors show increase andcool colors show decrease of spectral power relative to the baseline. Dashed vertical line shows the time of the second button-pressmarking themoment when themental image becamevivid. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

4 G.G. Knyazev et al. / International Journal of Psychophysiology xxx (2014) xxx–xxx

independent component patterns with known functional systems ornetworks. It compares the component's image with a template usingspatial correlation (Pearson's r) analysis. We used several templatesthat were generated using the Wake Forest Pick atlas toolbox (http://www.fmri.wfubmc.edu/). Posterior and anterior DMN, hippocampus,and WM network were considered potentially relevant to the study ofautobiographical memory. The anterior DMN hub (ADMN) templateincluded the superior frontal gyrus (BAs 8/9/10) and the anterior cingu-late cortex (BAs 11/32). The posterior DMN hub (PDMN) templateincluded the posterior cingulate, the precuneus (BA 7), and theoccipitoparietal junction (BA 39). The hippocampal/parahippocampalarea included BAs 20, 27, 28, 34, 35, and 36. The temporal template in-cluded BAs 21, 22, 37, and 38. TheWM template included the dorsolat-eral (BA 46) and the ventrolateral (BAs 44, 45, and 47) prefrontal cortex.It is generally accepted now that, besides the classical prefrontal areas,WM networks include a set of parietal regions, as well as content-specific areas, such as occipital cortex, auditory cortex, and motor/premotor areas (Owen et al., 2005; Zimmer, 2008). ParietalWMregions(e.g., BA 7) considerably overlap with the posterior DMN hub(e.g., Owen et al., 2005), therefore the parietal WM template was notused. The occipital template included the primary and secondary visualcortex (BAs 17, 18, and 19), the auditory cortex template included BAs41 and 42, and the motor/premotor template included respectivemotor, premotor, and somatosensory cortices (BAs 1 through 6). Moremethodological details of these methods can be found in Knyazevet al. (2011, 2012, 2013).

2.4.3. Statistical analysisWithin the TICA approach, mass-univariate testing (Friston, 2003;

Worsley et al., 1996) was used to reveal statistically significant changesof spectral power in the test interval relative to the baseline. Specifically,one-sample (across participants) T-test was independently fit at eachelement of time–frequencymatrix of ERSP values for each cluster. Inde-pendent samples T-testwas applied to reveal gender-related differencesin these effects. To reveal within-subject differences between trials withdifferent self-reported characteristics all trials for each participant weredivided into two groups, i.e., low and high values on respective charac-teristic (e.g., episodes, which were reported to arouse no or small posi-tive/negative emotion vs. episodes, which were accompanied by

Please cite this article as: Knyazev, G.G., et al., Oscillatory correlates of aut10.1016/j.ijpsycho.2014.12.006

moderate to strong positive/negative emotion). Because most partici-pants reported more positive and less negative emotion, it was decidedto retain in these analyses only those participantswhohad at least threeepisodes of each category, in order to make the statistical design morebalanced. Therefore, for the analysis of the effect of positive emotion,we had data for 46 participants, for negative emotion—37 participants,and for vividness—41 participant. Paired-samples T-test was applied toreveal statistically significant differences in ERSP values between thesetwo groups of episodes. Two-way ANOVAwas used to reveal significantinteractions of the within-subject effects with participant's gender. TheFalse Discovery Rate (FDR) correction was applied to correct for multi-ple comparisons (Holm, 1979).

In the case of SICA, two-level General Linear Model (GLM) approachsuggested by Calhoun et al. (2004) was used. First, a design matrix ofrelevant predictors was created according to a random effects GLManalysis (Friston et al., 1999). To reveal the general effect of recollection,the predictorwas defined as a boxcar functionwith zero valuesmarkingbaselines and unity values elsewhere. To reveal the effect of self-reported characteristics of each episode, the predictor was defined foreach participant separately as a boxcar functionwith values characteriz-ing the level of respective characteristic in each trial. A column of oneswas included in each design matrix in order to model the intercept.These design matrices were regressed on time courses of each compo-nent using the temporal sorting analysis in the GIFT software and nor-malized regression coefficients (‘beta’ weights) of the regressors wereused in the second-level analysis, which was performed using the‘stats on beta weights’ utility of the GIFT toolbox. One-sample T-testwas used to reveal whether beta weights were significantly differentfrom zero across the participants; independent samples T-test was ap-plied to find gender-related differences. FDR correction was applied tocorrect for multiple comparisons.

3. Results

3.1. Behavioral results

Across all participants, mean (SD) strength of positive emotion was3.5 (0.5), negative 1.9 (0.6), meaning that participants preferred to re-call emotionally positive episodes. Mean (SD) vividness was 4.3 (0.4);

obiographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/

Page 5

5G.G. Knyazev et al. / International Journal of Psychophysiology xxx (2014) xxx–xxx

hence, on average, participants deemed their recollections quite vivid.There were no gender differences in the mean ratings of emotion andvividness. In the between-subject domain, averaged over all episodes,positive emotion scores correlated negatively with averaged negativeemotion scores (r = −0.40, p = 0.006) and positively with averagedepisode vividness (r= 0.39, p= 0.008). Hence, participants who expe-rienced stronger positive emotions also tended to report less negativeemotions and deemed their recollections more vivid. To estimatewithin-subject (i.e., across the trials) correlations between psychomet-ric variables, they were calculated for each participant separately anda one-sample T-testwas applied to Fisher-Z-transformed correlation co-efficients in order to reveal significant after Bonferroni correctionacross-subjects correlations. Positive emotion scores correlated posi-tively with vividness (average r = 0.14, t(46) = 3.07, p = 0.004) andnegatively with negative emotion scores (r = −0.78, t(46) = −3.26,p = 0.002).

3.2. General effect of autobiographical memory

Channel-level one-sample T-test showed an increase of spectralpower in alpha and, to a lesser degree, in beta and a decrease of spectralpower in delta band. These effects were most pronounced after the sec-ond button-press (Fig. 1).

3.2.1. TICA resultsFifteen component clusters were created and one-sample T-test was

applied to reveal statistically significant across participants ERSP values.Two component clusters mostly showed a significant increase of spec-tral power in alpha band, whereas three other clusters mostly showeda decrease of power in delta band during the test interval relative tothe baseline (Fig. 2).

The increase of alpha powerwas observed in clusters,which showedthe highest spatial correlations with motor and ADMN templates; thedecrease of delta power was observed in clusters associated with WM,temporal, and motor area. Effect of gender was significant for the com-ponent cluster showing the highest spatial correlation with the leftmotor template. For this cluster, the increase of alpha power wasmore pronounced in males, whereas the decrease of delta power wasmore pronounced in females.

3.2.2. SICA resultsOne-sample T-test on beta weights showed significant effects for 9

components in delta, 1 component in theta, 13 components in alpha,and 2 components in beta band. In delta and theta bands, all significantbeta weights were negative, whereas in alpha and beta they were posi-tive. Note that negative and positiveweightsmean a decrease and an in-crease of power in the test interval relative to the baseline, respectively.No significant effects were found in gamma band. Fig. 3 presents topog-raphy and time-courses of components showing strongest effects(i.e., the highest T values) in the three frequency bands.

Note, that in time course plots, vertical lines mark the baselines,whereas points between these lines belong to the test interval in eachtrial (i.e., 10 trials, each of which consists of 10 one-second epochs). Itis evident that for delta and theta band components, there is mostly de-crease of power in the test interval relative to the baseline, whereas foralpha and beta band components the opposite pattern prevails. Spatialcorrelation analysis revealed that the largest number of componentsshowed strongest correlations with the posterior DMN (1 componentin delta, 1 in theta, 5 in alpha, and 1 in beta) and the occipital (2 compo-nents in delta and 5 in alpha) templates. Two components correlatedwith the WM template (1 in delta and 1 in alpha), four with themotor/premotor template (2 in delta and 2 in alpha), two with hippo-campal/parahippocampal template (1 in delta and 1 in beta), and onedelta component correlated with the anterior DMN template (Fig. 4).

In delta frequency band, effect of gender was significant for twocomponents showing the highest spatial correlations with the motor

Please cite this article as: Knyazev, G.G., et al., Oscillatory correlates of auto10.1016/j.ijpsycho.2014.12.006

and the right posterior DMN templates. In both cases, the decrease ofdelta power was more pronounced in females than in males.

3.3. Effect of psychometric variables

Table 1 summarizes the effects of psychometric variables. Bothmethods showed that positive emotion was associated with greaterdelta-band desynchronization; the difference between episodes withhigh and low positive emotion was higher in females than in males.Themain effect of positive emotionwas found in the precuneus, where-as the interaction with gender was found in the ACC and the right tem-poral lobe. Bothmethods showed that negative emotion was associatedwith a diminished alpha-band synchronization in cortical areas show-ing the highest spatial correlation with the PDMN. Additionally, TICAshowed an increase of delta power in the inferior frontal gyrus in epi-sodes with high negative emotion and its decrease in episodes withlow negative emotion. SICA additionally showed a diminished beta-band synchronization in the ACC in episodes with high negative emo-tion. There were no effects of gender for negative emotion. An increaseof theta power in episodes with high vividness and a decrease in epi-sodes with low vividness was found using the TICA approach in the su-perior temporal gyrus (the highest spatial correlation with the auditorytemplate). Both methods showed that the difference between episodeswith high and low vividnesswas greater in females than inmales in cor-tical areas showing the highest spatial correlation with the motortemplate.

4. Discussion

In this study,we aimed to compare EEG correlates of autobiographicalmemory with the existing fMRI evidence. For this reason, one of the twomethods of EEG data analysis that were used in this study (i.e., the SICAapproach) was designed so that it would maximally match the relevantfMRI studies, whereas the othermethod (i.e., the TICA approach)was de-signed following the tradition of EEG research. These two methods arebased on different assumptions (i.e., temporal vs. spatial independenceof reconstructed sources). Besides, they are used here in combinationwith different source localization techniques (i.e., equivalent dipolefitting vs. distributed source imaging). For these reasons, they are not ex-pected to produce analog results. Rather, they should be treated as com-plementary approaches, each of which has its own strengths andlimitations. Nevertheless, simulation studies show that in some generalrelatively simple cases the results of these methods are similar(Knyazev, 2013c; Peterson et al., 2000). In this study, results obtainedusing these two approaches are also in good agreement with eachother. Both methods show that autobiographical remembering is associ-ated with an increase of spectral power in alpha and beta and a decreasein delta band. These effects are observed in three functional cortical do-mains, which include the classical memory-related cortical areas, suchas hippocampus and frontal WM regions, content-specific areas, such asvisual, auditory, and motor/premotor cortex, and the DMN. In line withexisting fMRI evidence (e.g., Spreng et al., 2008), the posterior DMNhub showed the most prominent involvement, being represented byeight SICA components belonging to four frequency bands (see Fig. 4).In our previous studies, EEG correlates of spontaneous self-referentialthoughts were found in the alpha frequency band in these same corticalregions (Knyazev et al., 2011, 2012). Hence, the increase of alpha powerin posterior DMN observed in this study most probably reflects self-referential component of autobiographical memory.

In terms of the number of relevant components, the second place be-longs to the occipital network, in line with some studies highlighting therole visual perspective plays in autobiographical memory retrieval(e.g., Sutin and Robins, 2008). Activation of motor cortical areas duringautobiographical remembering is also frequently observed in fMRI stud-ies (e.g., Svoboda et al., 2006). Hence, on first impression, it appears thatbesides confirming fMRI evidence in EEG domain, these findings do not

biographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/

Page 6

Fig. 2. TICA results. ERSP and localization of dipoles for five component clusters that showed significant (one-sample T-test with FDR correction) changes of spectral power during therecollection of autobiographical episodes. A—middle frontal gyrus (BA 6), spatial correlation with the motor template = 0.45; B—medial frontal cortex (BA 8) spatial correlation withthe ADMN= 0.32; C—inferior frontal gyrus (BA 46) spatial correlation with theWM template = 0.46; D—superior temporal gyrus (BA 22) spatial correlation with the left temporal tem-plate = 0.33; E—middle frontal gyrus (BA 6), spatial correlation with the motor template = 0.54.

6 G.G. Knyazev et al. / International Journal of Psychophysiology xxx (2014) xxx–xxx

bring new information about brain activity during autobiographical re-membering. This is not the case, however. Although currently a relativelydirect correlation seems to be favored between population synaptic ac-tivity and fMRI signals, these signals are actually ‘blind’ as to the nature(i.e., inhibitory or excitatory) of this activity (Arthurs and Boniface,2002; Ekstrom, 2010; Logothetis, 2008). Relation of EEG oscillations toinhibition and excitation in the brain is also far from being clear, butmuch data have been accumulated showing that externally oriented be-havior is generally associated with a decrease of alpha power and an in-crease of lower frequency oscillations (see e.g., Klimesch, 1999). It isimportant to emphasize that these data relate to the so-called inducedoscillations, that is, oscillations that are not phase-locked to the stimulus

Please cite this article as: Knyazev, G.G., et al., Oscillatory correlates of aut10.1016/j.ijpsycho.2014.12.006

(Pfurtscheller, 2003). Increase of evoked (i.e., phase-locked to the stimu-lus) alpha oscillations has been shown upon presentation of emotional(e.g., negative facial expressions) stimuli (Güntekin and Basar, 2007).These data prompted an assumption that spontaneous and inducedalpha oscillations reflect inhibition (Klimesch, 1996; Pfurtscheller,2003). Later it has been suggested that alpha power increase participatesin the integration of relevant cortical areas and simultaneously inhibitsirrelevant ones, i.e., those areas that may supply an input, which is aptto interrupt the ongoing mental process (Klimesch et al., 2007). Thegrowing body of evidence gives a good ground for the suggestion thatsynchronization of internal mental processes and simultaneous inhibi-tion of movement and the perception of exteroceptive stimuli might be

obiographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/

Page 7

Fig. 3. SICA results. Topography of components showing strongest effects (i.e., the highest T values in the one-sample T-test on beta weights of regressors contrasting test interval with thebaseline) in delta, theta, alpha, and beta frequency bands.

7G.G. Knyazev et al. / International Journal of Psychophysiology xxx (2014) xxx–xxx

the primary correlate of alpha oscillations (Knyazev, 2013b; Knyazevet al., 2008, 2011). From such a standpoint, the observed increase ofalpha power in visual and motor cortical areas may actually reflect inhi-bition of these processes. That is, the increase of alpha power in the DMNand WM regions may be associated with integration of cortical areas,which are directly involved in the reconstruction of autobiographicalmemory episodes, whereas the increase of alpha power in visual andmotor cortices may reflect inhibition of irrelevant areas that may supplyan input, which is apt to interrupt the ongoingmental process (Klimeschet al., 2007). An alternative interpretationwould be that motor and visu-al cortices are activated because they play an important role in the recon-struction of a vivid mental image. Autobiographical remembering isoften expressed in terms of traveling back in time to relive the event(Tulving, 2002). The ability mentally to time travel to reexperiencepast events is thought to depend on frontal lobe structures (Wheeleret al., 1997) and the activity in context-specific, especially visual areas(Kosslyn et al., 2001), because visual perspective plays an importantrole in autobiographical memory retrieval (see e.g., Sutin and Robins,2008). Moreover, alpha enhancement has been repeatedly shown intasks involving visual imagination (e.g. Cooper et al., 2003, 2006). Ifsuch interpretation were true, however, one would expect to find a

Please cite this article as: Knyazev, G.G., et al., Oscillatory correlates of auto10.1016/j.ijpsycho.2014.12.006

positive correlation between the increase of alpha power in visual andmotor cortices and subjective ratings of memory vividness, which isnot the case in this study. Another possibility is that many thoughtshave to be actively suppressed (inhibited) during holding an autobio-graphical memory in mind and this is what is reflected in the alpha syn-chronization. From such a standpoint, it is difficult to explain, however,why alpha synchronization ismost prominent in DMN,which, accordingto fMRI findings, is involved in self-referential thought and autobio-graphical memory.

The interpretation of the observed decrease of delta power is lessstraightforward, mostly due to limited available information related tothis frequency band. Increase of delta power is traditionally associatedwith either pathological processes or widespread inhibition (seeKnyazev, 2012 for a review). In awake normally functioning brain, in-crease of delta power may be associated with attention, particularlythe detection of motivationally salient cues in the environment. In-crease of delta power is also frequently observed in states of urgentneeds, which are subjectively experienced as aversive and are accompa-nied by diminished release of dopamine in respective receptive fields.Alternatively, enhanced dopaminergic neurotransmission would causea decrease of delta power (see Knyazev, 2007, 2012 for reviews).

biographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/

Page 8

Fig. 4. SICA results. Topography of significant components in one-sample T-test on beta weights of regressors contrasting test interval with the baseline (FDR correction). Components aresorted based on the highest spatial correlation with the respective spatial template. The number below each map shows the size of spatial correlation with respective template.

8 G.G. Knyazev et al. / International Journal of Psychophysiology xxx (2014) xxx–xxx

Therefore, one may tentatively suggest that the observed in this studydecrease of delta power during autobiographical remembering mightbe associated with diminished attention to the environment and gener-ally positive emotional state. Indeed, behavioral results of this study arewell in linewith the evidence showing that normally functioningpeopleremember a greater number of positive personal events than negativeones. The intensity of emotions prompted by recall of positive events

Table 1Effect of psychometric variables.

Method Band Localization

Positive emotionTICA Delta Precuneus (BA 7)SICA Delta Precuneus (BA 7)

Effect of gender for positive emotionTICA Delta/theta ACC (BA 24)SICA Delta STG (BA 38)

Negative emotionTICA Delta/theta IFG (BA 46)TICA Alpha PCC (BA 31)SICA Alpha IPL (BA 40)SICA Beta ACC (BA 24)

VividnessTICA Theta STG (BA 41)

Effect of gender for vividnessTICA Delta/theta MFG (BA 6)SICA Delta Precentral (BA 6)SICA Theta Precentral (BA 4)

TICA—Temporal independent component analysis; SICA—Spatial independent component analyferior frontal gyrus; PCC—Posterior cingulate; IPL—Inferior parietal lobule; MFG—Medial frontaRight Temp—Right temporal template; WM—Working memory template; ERD—Event-related

a The difference between episodes with high and low positive emotion or vividness is largeb Beta weights of the regression of positive emotion or vividness scores on the component's

Please cite this article as: Knyazev, G.G., et al., Oscillatory correlates of aut10.1016/j.ijpsycho.2014.12.006

fades less rapidly over time than does the intensity of emotionsprompted by recall of negative events (Levine and Bluck, 2004;Walker et al., 1997, 2003). In line with this hypothesis, the decrease ofdelta power in PDMN was more pronounced in episodes associatedwith greater positive emotion. Moreover, in episodes accompanied bynegative emotion, an increase of delta power in WM regions was ob-served. Because most people's self-schemas are generally positive,

Template Effect

PDMN (r = 0.42) Greater ERD in high emotionPDMN (r = 0.58) Greater ERD in high emotion

ADMN (r = 0.45) Stronger effect in femalesa

Right temp (r = 0.42) Stronger effect in femalesb

WM (r = 0.37) ERD in low emotion, ERS in high emotionPDMN (r = 0.36) Greater ERS in low emotionPDMN (r = 0.48) Greater ERS in low emotionADMN (r = 0.25) Greater ERS in low emotion

Auditory (r = 0.54) Greater ERS in high vividness

Motor (r = 0.36) Stronger effect in femalesa

Motor (r = 0.50) Stronger effect in femalesb

Motor (r = 0.53) Stronger effect in femalesb

sis; BA—Brodmann area; ACC—Anterior cingulate; STG—Superior temporal gyrus; IFG—In-l gyrus; PDMN—Posterior default mode network; ADMN—Anterior default mode network;desynchronization; ERS—Event-related synchronization.r in females than in males.time course are larger in females than in males.

obiographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/

Page 9

9G.G. Knyazev et al. / International Journal of Psychophysiology xxx (2014) xxx–xxx

positive experiences may be more likely to be self-relevant and may bemore easily integrated into a person's conception of himself or herselfthan negative ones (Holland and Kensinger, 2010). That may explainwhy the increase of alpha power in PDMNwasmore pronounced in ep-isodes, which were not associated with negative emotion.

Interestingly, most gender-related effectswere observed in the deltafrequency band. In most cases, event-related decrease of delta powerwasmore pronounced in females than inmales, particularly in episodes,which were rated as highly vivid andwere accompanied by strong pos-itive emotions. This is generally in linewith substantial literature, whichindicates that women are more emotionally expressive than men areand show stronger psychophysiological responses to emotional stimuli(Knyazev et al., 2009, 2010; Kring andGordon, 1998; Orozco and Ehlers,1998).

Contrary to the traditional view and the huge amount of evidencelinking memory processes with theta oscillations, few effects havebeen observed in this frequency band. Autobiographical rememberingwas associated with a decrease of theta power in the PDMN. However,an increase of theta power in the right temporal lobe (thehighest spatialcorrelation with the auditory template) was observed in episodes withhigh vividness as compared to episodes with low vividness. The formereffect is in linewith simultaneous EEG-fMRI studies, which show a neg-ative association of theta with DMN activity (Meltzer et al., 2007;Mizuhara et al., 2004; Scheeringa et al., 2008; White et al., 2012). Thelatter effect is generally in line with the traditional view on the involve-ment of temporal-lobe theta oscillations inmemory processes.Whetherthis effect is specifically associatedwith theprocessing of auditory infor-mation or relates to a broader set of processes, which take place in thetemporal lobe during memory retrieval is an open question. In general,it could be hypothesized that memory retrieval was associated with anincrease of theta power in the temporal lobe and the degree of this in-crease may correlate with vividness of the recollection. However, thestudy design did not favor the detection of these initial stages of auto-biographical remembering. Judgments of appropriate vividness of recol-lection (i.e., the second button press) are metacognitive judgments thatoccur behaviorally after thememory has been fully formed based on thequality of information retrieved (Rubin et al., 2003). Therefore, the anal-ysis of activity that occurred over a period of ten seconds after the re-trieval has already occurred (i.e., during the process of maintainingrather than retrieving) more favored the detection of processes thatare relevant to self-referential thought and are associated with theDMN. One may argue that in view of the evidence linking mid-frontaltheta with working memory (Gevins et al., 1997; Jensen and Tesche,2002; Krause et al., 2000; Onton et al., 2005), an increase of theta oscil-lations during the test interval would be expected. However, frontaltheta power has been shown to increase in a wide range of cognitivetasks beyond working memory, such as mental arithmetic (Asadaet al., 1999; Inanaga, 1998; Inouye et al., 1994; Ishii et al., 1999;Lazarev, 1998; Smith et al., 1999), error detection (Luu et al., 2003,2004), and language comprehension (Bastiaansen et al., 2002; Haldet al., 2006). The common property of all these conditions is that theyare associated with a diminished activity of the DMN (Scheeringaet al., 2008).

This study has a number of limitations, which partly are inherent inEEG research and partly are the consequence of the chosen experimen-tal paradigm and approach to data analysis. It should be stressed thatcomparative to fMRI, EEGmethod has low spatial resolution. Themeth-od that was used in this study (i.e., combination of SICA and TICA withsource localization and identification of components with known func-tional networks bymeans of spatial correlation) gives only approximatelocalization of observed effects. Moreover, although all reported spatialcorrelationswere highly significant due to large number of voxels in im-ages, some of them are relatively low, particularly for deeply locatedsources, such as hippocampus (Pearson's r range = 0.18 to 0.54). InfMRI studies that used the same approach (i.e., spatialmatching of com-ponents with predetermined spatial template), higher correlations are

Please cite this article as: Knyazev, G.G., et al., Oscillatory correlates of auto10.1016/j.ijpsycho.2014.12.006

usually reported. For example, Harrison et al. (2008), using the samesoftware, reported correlations with DMN template ranging from 0.40to 0.60. All this prompts caution in interpretation of spatial localizationof observed effects. However, reasonable confidence could be derivedfrom the fact that results of two different approaches to data analysisare in relatively good agreement with each other and with results ofmany published fMRI studies of autobiographical memory. The choiceof having participants decide for themselves what episodes to recallmay be considered a limitation. Most participants in most trials pre-ferred to recall emotionally positive episodes and this fact limits the op-portunity to reveal effects associated with negative emotion. In thisstudy, we deliberately concentrated on the analysis of relatively lateand stable stages of autobiographical remembering, which mostly in-clude formation and maintenance rather than retrieval of memory(Conway et al., 2001, 2003). The study of earlier stages in the future re-search may help to reveal effects associated with retrieval of autobio-graphical memories and initial emotional reaction.

5. Conclusion

In this study, EEG correlates of autobiographical memory were in-vestigated using two different approaches to data analysis. Bothmethods gave comparable results. In terms of cortical localization of ob-served effects, these results are in line with existing fMRI evidence, butdue to different nature of EEG data, they allow for an interpretation,which is not usually considered in fMRI research. Autobiographical re-membering was associated with an increase of spectral power inalpha and beta and a decrease in delta band. These effects were ob-served in three functional cortical domains, including the classicalmemory-related areas, such as temporal lobe and frontal WM regions,the DMN, and content-specific areas, such as visual, auditory, andmotor cortex. The increase of alpha power was most prominent in theposterior DMN, followed by the visual and motor cortices. Taking intoaccount the well-known correlates of alpha activity (i.e., a positive cor-relation with DMN-related mental processes and a negative correlationwith movement and perception of exteroceptive stimuli), the mostplausible interpretation of this topography would be that during auto-biographical remembering, alpha oscillations are associatedwith activa-tion of DMN andwith simultaneous inhibition of irrelevant sensory andmotor areas. The decrease of delta power was observed in memory-related areas (WMand hippocampus), DMN, aswell as inmotor and oc-cipital regions. Given the data linking delta oscillations with attentionalprocesses (i.e., salience detection) and aversive states, effects in this fre-quency band could be tentatively interpreted as evidence of diminishedattention to the environment and generally positive emotional disposi-tion during autobiographical remembering. The latter supposition issupported by the fact that the decrease of delta power in DMN wasmore pronounced in episodes associatedwith greater positive emotion,whereas episodes associated with negative emotion were accompaniedby an increase of delta power. In line with existing evidence indicatingthatwomen show stronger psychophysiological responses to emotionalstimuli, gender-related effects were most evident in episodes, whichwere rated as highly vivid and were accompanied by strong positiveemotions. Summing up, this study shows that autobiographical remem-bering is associatedwith an increase of alpha and beta and a decrease ofdelta and theta oscillations, which are most prominent in the posteriorDMN hub. These findings are in line with the data linking DMN withalpha and beta oscillations and showing the leading role of this networkin autobiographical memory.

Acknowledgments

This study was supported by a grant of the Russian Foundation forBasic Research (RFBR), research project . 14-06-00039, and the RussianScientific Foundation (RSCF), research project No. 14-15-00202.

biographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/

Page 10

10 G.G. Knyazev et al. / International Journal of Psychophysiology xxx (2014) xxx–xxx

References

Annett, M., 1970. A classification of hand preference by association analysis. Br. J. Psychol.61, 303–321.

Arthurs, O.J., Boniface, S., 2002. Howwell dowe understand the neural origins of the fMRIBOLD signal? Trends Neurosci. 25, 27–31.

Asada, H., Fukuda, Y., Tsunoda, S., Yamaguchi, M., Tonoike, M., 1999. Frontal midline thetarhythms reflect alternative activation of prefrontal cortex and anterior cingulate cor-tex in humans. Neurosci. Lett. 274, 29–32.

Bastiaansen, M.C.M., van Berkum, J.J.A., Hagoort, P., 2002. Syntactic processing modulatesthe theta rhythm of the human EEG. NeuroImage 17, 1479–1492.

Brookes, M.J., Woolrich, M., Luckhoo, H., Price, D., Hale, J.R., Stephenson, M.C., Barnes, G.R.,Smith, S.M., Morris, P.G., 2011. Investigating the electrophysiological basis of restingstate networks using magnetoencephalography. Proc. Natl. Acad. Sci. U. S. A. 108,16783–16788.

Buzsaki, G., 2005. Theta rhythm of navigation: link between path integration and land-mark navigation, episodic and semantic memory. Hippocampus 15, 827–840.

Calhoun, V.D., Adali, T., Pearlson, G.D., Pekar, J.J., 2001. A method for making group infer-ences from functional MRI data using independent component analysis. Hum. BrainMapp. 14, 140–151.

Calhoun, V.D., Adali, T., Pekar, J.J., 2004. A method for comparing group fMRI data usingindependent component analysis: application to visual, motor and visuomotortasks. Magn. Reson. Imaging 22, 1181–1191.

Cannon, R.L., Baldwin, D.R., 2012. EEG current source density and the phenomenology ofthe default network. Clin. EEG Neurosci. 43, 257–267.

Conway, M.A., Turk, D.J., Miller, S.L., Logan, J., Nebes, R.D., Meltzer, C.C., Becker, J.T., 1999. Apositron emission tomography (PET) study of autobiographical memory retrieval.Memory 7, 679–702.

Conway, M.A., Pleydell-Pearce, C.W., Whitecross, S.E., 2001. The neuroanatomy of auto-biographical memory: a slow cortical potential study of autobiographical memory re-trieval. J. Mem. Lang. 46, 493–524.

Conway, M.A., Pleydell-Pearce, C.W., Whitecross, S.E., Sharpe, H., 2003. Neurophysiologi-cal correlates of memory for experienced and imagined events. Neuropsychologia 41,334–340.

Cooper, N.R., Croft, R.J., Dominey, S.J.J., Burgess, A.P., Gruzelier, J.H., 2003. Paradox lost? Ex-ploring the role of alpha oscillations during externally vs. internally directed atten-tion and the implications for idling and inhibition hypotheses. Int. J. Psychophysiol.47, 65–74.

Cooper, N.R., Burgess, A.P., Croft, R.J., Gruzelier, J.H., 2006. Investigating evoked and in-duced electroencephalogram activity in task-related alpha power increases duringan internally directed attention task. Neuroreport 17, 205–208.

Delorme, A., Makeig, S., 2004. EEGLAB: an open source toolbox for analysis of single-trialEEG dynamics including independent component analysis. J. Neurosci. Methods 134,9–21.

Düzel, E., Penny, W.D., Burgess, N., 2010. Brain oscillations and memory. Curr. Opin.Neurobiol. 20, 143–149.

Ekstrom, A., 2010. How and when the fMRI BOLD signal relates to underlying neural ac-tivity: the danger in dissociation. Brain Res. Rev. 62, 233–244.

Foster, B.L., Kaveh, A., Dastjerdi, M., Miller, K.J., Parvizi, J., 2013. Human retrosplenial cor-tex displays transient theta phase locking with medial temporal cortex prior to acti-vation during autobiographical memory retrieval. J. Neurosci. 33, 10439–10446.

Friston, K.J., 2003. Experimental design and statistical parametric mapping. In:Frackowiak, R., Friston, K., Frith, C., Dolan, R., Price, C. (Eds.), Human Brain Function,2nd edition Elsevier Press, London, pp. 599–632.

Friston, K.J., Holmes, A.P., Price, C.J., Buchel, C., Worsley, K.J., 1999. Multisubject fMRI stud-ies and conjunction analyses. NeuroImage 10, 385–396.

Gevins, A., Smith, M.E., McEvoy, L., Yu, D., 1997. High-resolution EEG mapping of corticalactivation related to working memory: effects of task difficulty, type of processing,and practice. Cereb. Cortex 7, 374–385.

Güntekin, B., Basar, E., 2007. Emotional face expressions are differentiated with brain os-cillations. Int. J. Psychophysiol. 64, 91–100.

Gusnard, D.A., Akbudak, E., Shulman, G.L., Raichle, M.E., 2001. Medial prefrontal cortexand self-referential mental activity: relation to a default mode of brain function.Proc. Natl. Acad. Sci. U. S. A. 98, 4259–4264.

Hald, L.A., Bastiaansen, M.C.M., Hagoort, P., 2006. EEG theta and gamma responses to se-mantic violations in online sentence processing. Brain Lang. 96, 90–105.

Harrison, B.J., Pujol, J., Lopez-Sola, M., Hernandez-Ribas, R., Deus, J., Ortiz, H., Soriano-Mas,C., Yucel, M., Pantelis, C., Cardoner, N., 2008. Consistency and functional specializationin the default mode brain network. Proc. Natl. Acad. Sci. U. S. A. 105, 9781–9786.

Hlinka, J., Alexakis, C., Diukova, A., Liddle, P.F., Auer, D.P., 2010. Slow EEG pattern predictsreduced intrinsic functional connectivity in the default mode network: an inter-subject analysis. NeuroImage 53, 239–246.

Holland, A.C., Kensinger, E.A., 2010. Emotion and autobiographical memory. Phys. LifeRev. 7, 88–131.

Holm, S., 1979. A simple sequentially rejective multiple test procedure. Scand. J. Stat. 6,65–70.

Inanaga, K., 1998. Frontal midline theta rhythm and mental activity. Psychiatry Clin.Neurosci. 52, 555–566.

Inouye, T., Shinosaki, K., Iyama, A., Matsumoto, Y., Toi, S., Ishihara, T., 1994. Potential flowof frontal midline theta-activity during a mental task in the human electroencephalo-gram. Neurosci. Lett. 169, 145–148.

Ishii, R., Shinosaki, K., Ukai, S., Inouye, T., Ishihara, T., Yoshimine, T., Hirabuki, N., Asada, H.,Kihara, T., Robinson, S.E., Takeda, M., 1999. Medial prefrontal cortex generates frontalmidline theta rhythm. Neuroreport 10, 675–679.

Jann, K., Dierks, T., Boesch, C., Kottlow, M., Strik, W., Koenig, T., 2009. BOLD correlates of EEGalpha phase-locking and the fMRI default mode network. NeuroImage 45, 903–916.

Please cite this article as: Knyazev, G.G., et al., Oscillatory correlates of aut10.1016/j.ijpsycho.2014.12.006

Jann, K., Kottlow, M., Dierks, T., Boesch, C., Koenig, T., 2010. Topographic electrophysiolog-ical signatures of fMRI resting state networks. PLoS ONE 5, e12945.

Jensen, O., Lisman, J.E., 2005. Hippocampal sequence-encoding driven by a cortical multi-item working memory buffer. Trends Neurosci. 28, 67–72.

Jensen, O., Tesche, C.D., 2002. Frontal theta activity in humans increases with memoryload in a working memory task. Eur. J. Neurosci. 15, 1395–1399.

Kahana, M.J., Seelig, D., Madsen, J.R., 2001. Theta returns. Curr. Opin. Neurobiol. 11, 739–744.Keil, A., Debener, S., Gratton, G., Junghöfer, M., Kappenman, E.S., Luck, S.J., Luu, P., Miller,

G.A., Yee, C.M., 2014. Committee report: publication guidelines and recommenda-tions for studies using electroencephalography and magnetoencephalography. Psy-chophysiology 51, 1–21.

Klimesch, W., 1996. Memory processes, brain oscillations and EEG synchronization. Int.J. Psychophysiol. 24, 61–100.

Klimesch,W., 1999. EEG alpha and theta oscillations reflect cognitive andmemory perfor-mance: a review and analysis. Brain Res. Rev. 29, 169–195.

Klimesch, W., Sauseng, P., Hanslmayr, S., 2007. EEG alpha oscillations: the inhibition-timing hypothesis. Brain Res. Rev. 53, 63–88.

Klimesch, W., Freunberger, R., Sauseng, P., 2010. Oscillatory mechanisms of process bind-ing in memory. Neurosci. Biobehav. Rev. 34, 1002–1014.

Klinger, E., Gregoire, K.C., Barta, S.G., 1973. Physiological correlates of mental activity: eyemovements, alpha, and heart rate during imagining, suppression, concentration,search, and choice. Psychophysiology 10, 471–477.

Knyazev, G.G., 2007. Motivation, emotion, and their inhibitory control mirrored in brainoscillations. Neurosci. Biobehav. Rev. 31, 377–395.

Knyazev, G.G., 2012. EEG delta oscillations as a correlate of basic homeostatic andmotiva-tional processes. Neurosci. Biobehav. Rev. 36, 677–695.

Knyazev, G.G., 2013a. Extraversion and anterior vs. posterior DMN activity during self-referential thoughts. Front. Hum. Neurosci. 6, 348.

Knyazev, G.G., 2013b. EEG correlates of self-referential processing. Front. Hum.Neurosci. 7, 264.

Knyazev, G.G., 2013c. Comparison of spatial and temporal independent component anal-yses of electroencephalographic data: a simulation study. Clin. Neurophysiol. 124,1557–1569.

Knyazev, G.G., Bocharov, A.V., Levin, E.A., Savostyanov, A.N., Slobodskoj-Plusnin, J.Y., 2008.Anxiety and oscillatory responses to emotional facial expressions. Brain Res. 1227,174–188.

Knyazev, G.G., Bocharov, A.V., Slobodskoj-Plusnin, J.Y., 2009. Hostility- and gender-relateddifferences in oscillatory responses to emotional facial expressions. Aggress. Behav.35, 502–513.

Knyazev, G.G., Slobodskoi-Plyusnin, Y.Y., Bocharov, A.V., 2010. Gender differences in im-plicit and explicit processing of emotional facial expressions as revealed by event-related theta synchronization. Emotion 10, 678–687.

Knyazev, G.G., Slobodskoj-Plusnin, J.Y., Bocharov, A.V., Pylkova, L.V., 2011. The defaultmode network and EEG alpha oscillations: an independent component analysis.Brain Res. 1402, 67–79.

Knyazev, G.G., Savostyanov, A.N., Volf, N.V., Liou, M., Bocharov, A.V., 2012. EEG correlatesof spontaneous self-referential thoughts: a cross-cultural study. Int. J. Psychophysiol.86, 173–181.

Knyazev, G.G., Slobodskoj-Plusnin, J.Y., Bocharov, A.V., Pylkova, L.V., 2013. Cortical oscilla-tory dynamics in a social interaction model. Behav. Brain Res. 241, 70–79.

Kosslyn, S.M., Ganis, G., Thompson, W.L., 2001. Neural foundations of imagery. Nat. Rev.Neurosci. 2, 635–642.

Krause, C.M., Sillanmaki, L., Koivisto, M., Saarela, C., Haggqvist, A., Laine, M., Hamalainen,H., 2000. The effects of memory load on event-related EEG desynchronization andsynchronization. Clin. Neurophysiol. 111, 2071–2078.

Kring, A.M., Gordon, A.H., 1998. Sex differences in emotion: expression, experience, andphysiology. J. Personal. Soc. Psychol. 74, 686–703.

Laird, A.R., Eickhoff, S.B., Li, K., Robin, D.A., Glahn, D.C., Fox, P.T., 2009. Investigating thefunctional heterogeneity of the default mode network using coordinate-basedmeta-analytic modeling. J. Neurosci. 29, 14496–14505.

Lazarev, V.V., 1998. On the intercorrelation of some frequency and amplitude parametersof the human EEG and its functional significance. Communication I: multidimension-al neurodynamic organization of functional states of the brain during intellectual,perceptive and motor activity in normal subjects. Int. J. Psychophysiol. 28, 77–98.

Leech, R., Kamourieh, S., Beckmann, C.F., Sharp, D.J., 2011. Fractionating the default modenetwork: distinct contributions of the ventral and dorsal posterior cingulate cortex tocognitive control. J. Neurosci. 31, 3217–3224.

Levine, L.J., Bluck, S., 2004. Painting with broad strokes: happiness and the malleability ofevent memory. Cogn. Emot. 18, 559–574.

Li, Y., Adali, T., Calhoun, V.D., 2007. Estimating the number of independent componentsfor functional magnetic resonance imaging data. Hum. Brain Mapp. 28, 1251–1266.

Logothetis, N.K., 2008. What we can do and what we cannot do with fMRI. Nature 453,869–878.

Luu, P., Tucker, D.M., Derryberry, D., Reed, M., Poulsen, C., 2003. Electrophysiological re-sponses to errors and feedback in the process of action regulation. Psychol. Sci. 14, 47–53.

Luu, P., Tucker, D.M., Makeig, S., 2004. Frontal midline theta and the error-related negativ-ity: neurophysiological mechanisms of action regulation. Clin. Neurophysiol. 115,1821–1835.

Makeig, S., 1993. Auditory event-related dynamics of the EEG spectrum and effects of ex-posure to tones. Electroencephalogr. Clin. Neurophysiol. 86, 283–293.

Makeig, S., Delorme, A., Westerfield, M., Jung, T.P., Townsend, J., Courchesne, E.,Sejnowski, T.J., 2004. Electroencephalographic brain dynamics following manuallyresponded visual targets. PLoS Biol. 2, e176.

Mantini, D., Perrucci, M.G., Del Gratta, D., Romani, G.L., Corbetta, M., 2007. Electrophysio-logical signatures of resting state networks in the human brain. Proc. Natl. Acad. Sci.U. S. A. 104, 13170–13175.

obiographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/

Page 11

11G.G. Knyazev et al. / International Journal of Psychophysiology xxx (2014) xxx–xxx

Meltzer, J.A., Negishi, M., Mayes, L.C., Constable, R.T., 2007. Individual differences in EEGtheta and alpha dynamics during working memory correlate with fMRI responsesacross participants. Clin. Neurophysiol. 118, 2419–2436.

Mizuhara, H., Wang, L.Q., Kobayashi, K., Yamaguchi, Y., 2004. A long-range cortical net-work emerging with theta oscillation in a mental task. NeuroReport 15, 1233–1238.

Mo, J., Liu, Y., Huang, H., Ding, M., 2013. Coupling between visual alpha oscillations anddefault mode activity. NeuroImage 68, 112–118.

Moscovitch, M., Rosenbaum, R.S., Gilboa, A., Addis, D.R., Westmacott, R., Grady, C.,McAndrews, M.P., Levine, B., Black, S.E., Winocur, G., Nadel, L., 2005. Functional neu-roanatomy of remote episodic, semantic and spatial memory: a unified account basedon multiple trace theory. J. Anat. 207, 35–66.

Nyhus, E., Curran, T., 2010. Functional role of gamma and theta oscillations in episodicmemory. Neurosci. Biobehav. Rev. 34, 1023–1035.

Onton, J., Delorme, A., Makeig, S., 2005. Frontal midline EEG dynamics during workingmemory. NeuroImage 27, 341–356.

Onton, J., Westerfield, M., Townsend, J., Makeig, S., 2006. Imaging human EEG dynamicsusing independent component analysis. Neurosci. Biobehav. Rev. 30, 808–822.

Orozco, S., Ehlers, C.L., 1998. Gender differences in electrophysiological responses to facialstimuli. Biol. Psychiatry 44, 281–289.

Owen, A.M., McMillan, K.M., Laird, A.R., Bullmore, E., 2005. N-back working memory par-adigm: a meta-analysis of normative functional neuroimaging studies. Hum. BrainMapp. 25, 46–59.

Pascual-Marqui, R.D., 2002. Standardized low-resolution brain electromagnetic tomogra-phy (sLORETA): technical details. Methods Find. Exp. Clin. Pharmacol. 24 (Suppl. D),5–12.

Peterson, K.S., Hansen, L.K., Kolenda, T., Rostrup, E., Strother, S.C., 2000. On the indepen-dent components of functional neuroimages. In: Pajunen, P., Karhunen, J. (Eds.), Pro-ceedings ICA2000, Otamedia, Helsinki, pp. 615–620.

Pfurtscheller, G., 2003. Induced oscillations in the alpha band: functional meaning.Epilepsia 44 (Suppl. 12), 2–8.

Raichle, M.E., Gusnard, D.A., 2005. Intrinsic brain activity sets the stage for expression ofmotivated behavior. J. Comp. Neurol. 493, 167–176.

Raichle, M.E., MacLeod, A.M., Snyder, A.Z., Powers, W.J., Gusnard, D.A., Shulman, G.L.,2001. A default mode of brain function. Proc. Natl. Acad. Sci. U. S. A. 98, 676–682.

Ramkumar, P., Parkkonen, L., Hari, R., Hyvärinen, A., 2012. Characterization ofneuromagnetic brain rhythms over time scales of minutes using spatial independentcomponent analysis. Hum. Brain Mapp. 33, 1648–1662.

Ramkumar, P., Parkkonen, L., Hyvärinen, A., 2014. Group-level spatial independent com-ponent analysis of Fourier envelopes of resting-state MEG data. NeuroImage 86,480–491.

Ray, W.J., Cole, H.W., 1985a. EEG activity during cognitive processing: influence of atten-tional factors. Int. J. Psychophysiol. 3, 43–48.

Ray, W.J., Cole, H.W., 1985b. EEG alpha activity reflects attentional demands, and beta ac-tivity reflects emotional and cognitive processes. Science 228, 750–752.

Ros, T., Théberge, J., Frewen, P.A., Kluetsch, R., Densmore, M., Calhoun, V.D., Lanius, R.A.,2013. Mind over chatter: plastic up-regulation of the fMRI salience network directlyafter EEG neurofeedback. NeuroImage 65, 324–335.

Rubin, D.C., Schrauf, R.W., Greenberg, D.L., 2003. Belief and recollection of autobiograph-ical memories. Mem. Cognit. 31, 887–901.

Please cite this article as: Knyazev, G.G., et al., Oscillatory correlates of auto10.1016/j.ijpsycho.2014.12.006

Sadaghiani, S., Scheeringa, R., Lehongre, K., Morillon, B., Giraud, A.L., Kleinschmidt, A.,2010. Intrinsic connectivity networks, alpha oscillations, and tonic alertness: a simul-taneous electroencephalography/functional magnetic resonance imaging study.J. Neurosci. 30, 10243–10250.

Sadaghiani, S., Scheeringa, R., Lehongre, K., Morillon, B., Giraud, A.L., D'Esposito, M.,Kleinschmidt, A., 2012. Alpha-band phase synchrony is related to activity in thefronto-parietal adaptive control network. J. Neurosci. 32, 14305–14310.

Scheeringa, R., Bastiaansen, M.C.M., Petersson, K.M., Oostenveld, R., Norris, D.G., Hagoort,P., 2008. Frontal theta EEG activity correlates negatively with the default mode net-work in resting state. Int. J. Psychophysiol. 67, 242–251.

Sestieri, C., Corbetta, M., Romani, G.L., Shulman, G.L., 2011. Episodic memory retrieval, pa-rietal cortex, and the default mode network: functional and topographic analyses.J. Neurosci. 31, 4407–4420.

Smith, M.E., McEvoy, L.K., Gevins, A., 1999. Neurophysiological indices of strategy devel-opment and skill acquisition. Brain Res. Cogn. Brain Res. 7, 389–404.

Spreng, R.N., Mar, R.A., Kim, A.S.N., 2008. The common neural basis of autobiographicalmemory, prospection, navigation, theory of mind, and the default mode: a quantita-tive meta-analysis. J. Cogn. Neurosci. 21, 489–510.

Stella, F., Treves, A., 2011. Associativememory storage and retrieval: Involvement of thetaoscillations in hippocampal information processing. Neural Plast. 2011, 683961.

Sutin, A.R., Robins, R.W., 2008. When the “I” looks at the “Me”: autobiographical memory,visual perspective, and the self. Conscious. Cogn. 17, 1386–1397.

Svoboda, E., McKinnon, M.C., Levine, B., 2006. The functional neuroanatomy of autobio-graphical memory: a meta-analysis. Neuropsychologia 44, 2189–2208.

Tulving, E., 2002. Episodic memory: from mind to brain. Annu. Rev. Psychol. 53, 1–25.Vertes, R.P., 2005. Hippocampal theta rhythm: a tag for short-term memory. Hippocam-

pus 15, 923–935.Walker, W.R., Vogl, R.J., Thompson, C.P., 1997. Autobiographical memory: unpleasantness

fades faster than pleasantness over time. Appl. Cogn. Psychol. 11, 399–413.Walker, W.R., Skowronski, J.J., Thompson, C.P., 2003. Life is pleasant—and memory helps

keep it that way! Rev. Gen. Psychol. 7, 203–210.Wang, L., Saalmann, Y.B., Pinsk, M.A., Arcaro, M.J., Kastner, S., 2012. Electrophysiological

low-frequency coherence and cross-frequency coupling contribute to BOLD connec-tivity. Neuron 76, 1010–1020.

Wheeler, M.A., Stuss, D.T., Tulving, E., 1997. Toward a theory of episodic memory: thefrontal lobes and autonoetic consciousness. Psychol. Bull. 121, 331–354.

White, T.P., Jansen, M., Doege, K., Mullinger, K.J., Park, S.B., Liddle, E.B., Gowland, P.A.,Francis, S.T., Bowtell, R., Liddle, P.F., 2012. Theta power during encoding predictssubsequent-memory performance and default mode network deactivation. Hum.Brain Mapp. 34, 2929–2943.

Worsley, K.J., Marrett, S., Neelin, P., Vandal, A.C., Friston, K.J., Evans, A.C., 1996. A unifiedstatistical approach or determining significant signals in images of cerebral activation.Hum. Brain Mapp. 4, 58–73.

Wu, L., Eichele, T., Calhoun, V.D., 2010. Reactivity of hemodynamic responses and func-tional connectivity to different states of alpha synchrony: a concurrent EEG-fMRIstudy. NeuroImage 52, 1252–1260.

Zimmer, H.D., 2008. Visual and spatial working memory: from boxes to networks.Neurosci. Biobehav. Rev. 32, 1373–1395.

biographical memory, Int. J. Psychophysiol. (2014), http://dx.doi.org/