RESEARCH REPORT

Modification of EEG power spectra and EEG connectivityin autobiographical memory: a sLORETA study

Claudio Imperatori • Riccardo Brunetti • Benedetto Farina •

Anna Maria Speranza • Anna Losurdo • Elisa Testani •

Anna Contardi • Giacomo Della Marca

Received: 3 June 2013 / Accepted: 24 February 2014

� Marta Olivetti Belardinelli and Springer-Verlag Berlin Heidelberg 2014

Abstract The aim of the present study was to explore the

modifications of scalp EEG power spectra and EEG con-

nectivity during the autobiographical memory test (AM-T)

and during the retrieval of an autobiographical event (the

high school final examination, Task 2). Seventeen healthy

volunteers were enrolled (9 women and 8 men, mean age

23.4 ± 2.8 years, range 19–30). EEG was recorded at

baseline and while performing the autobiographical mem-

ory (AM) tasks, by means of 19 surface electrodes and a

nasopharyngeal electrode. EEG analysis was conducted by

means of the standardized LOw Resolution Electric

Tomography (sLORETA) software. Power spectra and

lagged EEG coherence were compared between EEG

acquired during the memory tasks and baseline recording.

The frequency bands considered were as follows: delta

(0.5–4 Hz); theta (4.5–7.5 Hz); alpha (8–12.5 Hz); beta1

(13–17.5 Hz); beta2 (18–30 Hz); gamma (30.5–60 Hz).

During AM-T, we observed a significant delta power

increase in left frontal and midline cortices (T = 3.554;

p \ 0.05) and increased EEG connectivity in delta band in

prefrontal, temporal, parietal, and occipital areas, and for

gamma bands in the left temporo-parietal regions

(T = 4.154; p \ 0.05). In Task 2, we measured an

increased power in the gamma band located in the left

posterior midline areas (T = 3.960; p \ 0.05) and a sig-

nificant increase in delta band connectivity in the pre-

frontal, temporal, parietal, and occipital areas, and in the

gamma band involving right temporo-parietal areas

(T = 4.579; p \ 0.05). These results indicate that AM

retrieval engages in a complex network which is mediated

by both low- (delta) and high-frequency (gamma) EEG

bands.

Keywords Autobiographical memory � sLORETA � EEG

power spectra � EEG connectivity

Introduction

Autobiographical memory (AM) is a complex form of

explicit memory referring to the ability to remember events

from one’s own life, and it is believed to be a dynamic

integration between episodic memory (EM) and semantic

memory (SM) (Cabeza and St Jacques 2007; Levine et al.

2004). EM is defined as a system characterized by an au-

tonoetic consciousness, reflecting the ability to remember

unique past events together with their associated contextual

details (Tulving 1985, 1987). In contrast, SM is charac-

terized by noetic consciousness, and it refers to the

knowledge of facts about the world and about our life

(Tulving 1985, 1987).

Autobiographical memories are characterized by both

elements of these memory systems such as vivid and

emotional images (EM) of people during a particular event

(i.e., high school final examination) with the knowledge

about concepts and facts related to this special day (SM).

Retrieving autobiographical information involves two dif-

ferent phases: event construction and elaboration (Conway

et al. 2001, 2003; Daselaar et al. 2008; Holland et al. 2011).

C. Imperatori � R. Brunetti � B. Farina (&) � A. Contardi

Department of Human Science, European University of Rome,

Rome, Italy

e-mail: benfarina@gmail.com

A. M. Speranza

Department of Dynamic and Clinical Psychology, Sapienza

University, Rome, Italy

A. Losurdo � E. Testani � G. Della Marca

Department of Neurosciences, Catholic University, Rome, Italy

123

Cogn Process

DOI 10.1007/s10339-014-0605-5

The first one is characterized by a dynamic and inferential

process based on an initial cue, and by a subsequent con-

tinuous monitoring and evaluation of the search results; the

second phase is characterized by the emerging of different

aspects and details about the autobiographical recall kept in

mind (Holland et al. 2011).

The complexity of the AM is reflected by its neurobio-

logical underpinnings. The study of AM neural substrates

has received significant attention in the last decade, and

recent neuroimaging investigations (Cabeza and St Jacques

2007; Holland et al. 2011; Rugg and Vilberg 2013; Svo-

boda et al. 2006) offered the possibility to map a relatively

consistent cooperation between different brain areas, pre-

dominantly left-lateralized, involving in particular the

prefrontal cortex (PFC) and different structures of the

mesial temporal lobe (MTL).

Cabeza and St Jacques (2007), in a review of functional

neuroimaging studies of AM, proposed a complex brain

network in which ‘memory search process,’ ‘monitoring

phase,’ and ‘self-referential process’ seem to involve,

respectively, left lateral PFC, medial PFC, and ventrome-

dial PFC. Moreover, the hippocampus and the retrosplenial

cortex are involved in the recollection during event elab-

oration process, whereas the amygdala, the occipital area,

and the cuneus/precuneus region are, respectively,

involved in emotional processing and in visual imagery

(Cabeza and St Jacques 2007).

Several electrophysiology studies confirmed the

involvement of PFC and MTL in AM. EEG and event-

related potentials (ERP) studies documented modifications

in brain electrical activity during the different phases of

AM retrieval. Conway et al. (Conway et al. 2001, 2003)

documented changes in slow cortical potentials during

construction and event elaboration phases: the first is

characterized by enhanced EEG activity in the left frontal

lobe, whereas the second is characterized by enhanced

EEG activity in posterior temporal and occipital lobes.

Furthermore, Steinvorth et al. (2010), using intracranial

ERP recordings in a single case, reported an increase in

gamma, theta, and delta frequency bands in the left en-

torhinal cortex: gamma was predominant in superficial

layers (which project to the hippocampus) during the pre-

sentation of the memory cue, whereas theta and delta were

prolonged and dominant in deep layers (which project to

neocortices) during memory retrieval. These authors also

observed that this last activation pattern was exclusive for

AM and was not observed in SM tasks (Steinvorth et al.

2010). Therefore, it has been proposed that changes in low-

frequency bands could endorse a long-range cortical

interaction within complex and spread brain networks

(Steinvorth et al. 2010; Toth et al. 2012). It has also been

proposed that the encoding and maintenance of memory

traces in MTL rely on the crucial modulation of brain

activity in theta and gamma frequencies (Duzel et al. 2010;

Steinvorth et al. 2010).

The aim of the present study was to explore the modi-

fications of scalp EEG power spectra and EEG connectivity

during AM tasks. For these reason, we used the AM test

(AM-T) (Williams and Broadbent 1986), one of the most

common and validated tests used to study AM both clini-

cally and experimentally (Van Vreeswijk and De Wilde

2004; Williams et al. 2007). Furthermore, in order to

explore the AM network organization in a more ecological

situation, we also explored the EEG modifications during

the retrieval of a single autobiographical event (namely, the

high school final examination, Task 2), based on concrete

questions (‘what,’ ‘where,’ ‘who’).

In order to detect modifications of EEG frequencies, and

their topographic distribution, we used the standardized

LOw Resolution brain Electric Tomography (sLORETA)

software, a validated method for localizing the electric

activity in the brain based on multichannel surface EEG

recordings (Pascual-Marqui et al. 1994). Furthermore, a

nasopharyngeal (NP) electrode was introduced via the

nostril and positioned with the tip touching the posterior

pharyngeal wall. This electrode is known to record EEG

activity originating from MTL structures (Zijlmans et al.

2008); in this way, we were able to collect the signal as

close as possible to the MTL and achieve a more accurate

source reconstruction for that region.

Materials and methods

Seventeen healthy volunteers were enrolled for the exper-

iment, 9 women and 8 men. Mean age was

23.4 ± 2.8 years (age range 19–30 years). The only

inclusion criterion was the consent to participate. Exclusion

criteria were as follows: left handedness; history of medi-

cal, psychiatric, and neurologic diseases; head trauma;

assumption of central nervous system active drugs in the

3 weeks before the study; presence of EEG abnormalities

at the baseline recording. The research was approved by

the Catholic University’s and Universita Europea’s ethics

review boards. Both the ethics review boards issued a

formal written waiver of informed consent because the

research involved no more than minimal risk, and the data

were analyzed anonymously. All subjects gave their writ-

ten informed consent to participate.

AM tasks

After electrode montage, the subjects were invited to sit in

a comfortable armchair, with eyes closed, in a quiet, semi-

dark silent room for a 5-min resting EEG recording

(baseline condition). Afterward, the subjects were

Cogn Process

123

instructed to perform the AM-T (Williams and Broadbent

1986): participants were asked to remain silent with their

eyes closed and to recall several autobiographical episodic

memories linked to 10 cue words. The cue words were read

out loud by a researcher (B.F.): five positive words (happy,

proud, faithful, tender, friendly) and five negative words

(tired, ashamed, painful, sad, weakness) translated and

adapted to Italian from Moradi et al. (2008). The AM-T

lasted 5 min.

Upon completing, the AM-T subjects were instructed to

remain seated with their eyes closed for 5 min before

starting the second task. In Task 2, participants were

instructed to recall and concentrate on the autobiographical

event of their high school exit examination. Subjects were

asked to remember as many details as possible about that

experience (e.g., people, emotions, venue, etc.). The sub-

ject was requested to perform the memory task for 5 min.

At the end of this session, the subjects were asked whether

they were able to successfully perform the task.

EEG recordings

Continuous EEG recordings were performed at baseline

and during the two AM tasks. EEG was recorded by means

of a Micromed System Plus digital EEGraph (Micromed�S.p.A., Mogliano Veneto, TV, Italy). EEG montage

included 19 standard scalp leads positioned according to

the 10–20 system (recording sites: Fp1, Fp2, F7, F3, Fz, F4,

F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, O1, O2), EOG,

and EKG. Moreover, a NP electrode was used to record

EEG activity in proximity of MTL structures. NP electrode

was introduced via the nostril and positioned with the tip

touching the posterior pharyngeal wall. The reference

electrodes were placed on the linked mastoids. Impedances

were kept below 5 KX before starting the recording and

checked again at the end. In particular, impedances of the

mastoids reference electrodes were checked to be identical.

Sampling frequency was 256 Hz; A/D conversion was

made at 16 bit; preamplifiers amplitude range was

±3,200 lV and low-frequency pre-filters were set at

0.15 Hz. The following band-pass filters were used:

HFF = 0.2 Hz; LFF = 128 Hz. The line noise (in Italy:

50 Hz) was removed by using a 50 Hz notch filter. Offline

artifact rejection (eye movements, blinks, muscular acti-

vations, or movement artifacts) was performed visually on

the raw EEG trace, by posing a marker at the onset of the

artifact signal and a further marker at the end of the artifact.

Successively, the artifact segment (that is, the EEG signal

interval included between the two markers) was deleted,

and this cancellation involved all the EEG traces acquired

within that interval. In this way, all the EEG intervals

characterized by the presence of artifacts were excluded

from the analysis. After artifact rejection, the remaining

EEG intervals were exported into American Standard Code

for Information Interchange (ASCII) files and imported

into the sLORETA software. We analyzed segments of

EEG recorded in baseline (BL) state and during tasks.

Since the duration of both tasks (AM-T and Task 2) and BL

was about 5 min, we decided to analyze at least 2 min of

artifact-free recording (not necessarily consecutive) for

each of the three condition (baseline, AM-T, Task 2), in all

subjects. The average time analyzed was 146 ± 19 s. This

procedure has been already used to investigate modifica-

tions of EEG power spectra during a working memory task

(N-back) (Imperatori et al. 2013), and in order to assess

EEG connectivity and EEG power spectra in dissociative

disorders (Farina et al. 2013). All EEG analyses were

performed by means of the sLORETA software (Pascual-

Marqui et al. 1994).

Frequency analysis

EEG frequency analysis was performed by means of fast

Fourier transform algorithm, with a 2-s interval on the EEG

signal, in all scalp locations. The following frequency

bands were considered: delta (0.5–4 Hz); theta

(4.5–7.5 Hz); alpha (8–12.5 Hz); beta1 (13–17.5 Hz);

beta2 (18–30 Hz); gamma (30.5–60 Hz). For frequency

analysis, monopolar EEG traces (each electrode referred to

joint mastoids) were used. Topographic sources of EEG

activities were determined using the sLORETA software.

The sLORETA software computes the current distribution

throughout the brain volume. In order to find a solution for

the 3-dimensional distribution of the EEG signal, the

sLORETA method assumes that neighboring neurons are

simultaneously and synchronously activated. This

assumption rests on evidence from single-cell recordings in

the brain that shows strong synchronization of adjacent

neurons (Kreiter and Singer 1992; Murphy et al. 1992).

The computational task is to select the smoothest of all

possible 3-dimensional current distributions, a common

procedure in signal processing (Grave de Peralta-Menen-

dez and Gonzalez-Andino 1998; Grave de Peralta Me-

nendez et al. 2000). The result is a true 3-dimensional

tomography, in which the localization of brain signals is

preserved with a low amount of dispersion (Pascual-Mar-

qui et al. 1994).

Connectivity analysis

The connectivity analysis was performed by the computa-

tion of lagged coherence. This approach allows to better

evaluate ‘true’ connectivity. Two measures of coherence

can be calculated between EEG signals: ‘lagged coherence’

Cogn Process

123

and ‘instantaneous coherence.’ These measures evaluate,

respectively, the contribution of potentials generated by

local neural networks and of potentials due to volume

conduction. The lagged coherence is a much more appro-

priate measure of electrophysiological connectivity,

because it removes the confounding effect of instantaneous

dependence due to volume conduction and low spatial

resolution (Pascual-Marqui 2007a). Furthermore, the lag-

ged component is purely physiological and affected mini-

mally by low spatial resolution, which affects the

instantaneous component (Pascual-Marqui et al. 2011).

However, other techniques such as the imaginary

coherence have been previously developed (Nolte et al.

2004), it is important to note that when there exists a

lagged connection, the imaginary part of the coherence

fails to detect it by tending to zero if the instantaneous

component is large. This is not the case for the lagged

coherence which asymptotically tends to a nonzero value,

detecting the presence of a physiological lagged connection

(Pascual-Marqui et al. 2011). For these reasons in the

present study, lagged coherence was calculated. The

sLORETA software computes instantaneous coherence

qx�y(x), in the case of univariate series, by the formula

(Pascual-Marqui 2007b):

Fig. 1 3D sLORETA

representation of the scalp

electrodes placement

Cogn Process

123

q2x:yðxÞ ¼

½ReðSyxxÞ�SyxxSxxx

whereas the lagged coherence qx$y(x) is calculated by the

formula (Pascual-Marqui 2007b):

q2x$yðxÞ ¼ 1� exp½�Fx$yðxÞ�

¼ 1

�

Syyx Syxx

Sxyx Sxxx

� ����������

Syyx O

OT Sxxx

� ���������

� �

ReSyyx Syxx

Sxyx Sxxx

� ����������

ReSyyx O

OT Sxxx

� ���������

� �

In this formula, ‘x’ is the discrete frequency considered,

‘Re’ indicates the real part of an element; Sxxx, Syyx, Sxyx,

and Syxx denote complex valued covariance matrices.

Fx$y(x) is the lagged linear dependence, ‘O’ is a matrix of

zeros, and the superscript ‘T’ stands for ‘transposed.’

The EEG coherence analysis was performed on the same

blocks of EEG tracings used for power spectra analysis.

Coherence values were computed for each frequency band

(delta, theta, alpha, beta, gamma), in the frequency range of

0.5–60 Hz. In order to evaluate the modifications of con-

nectivity, 20 region of interests (ROIs) were defined: one in

the temporal mesial region, placed on the midline, corre-

sponding to the site of the NP electrode, and 19 for the

scalp (one for each scalp electrode). In Fig. 1, we reported

the sLORETA representation of the scalp electrodes.

We chose the ‘single nearest voxel’ option: in this way,

each ROI consisted of a single voxel, the one closest to

each seed. Then, the sLORETA computed the coherence

values between all these ROIs (total 20 9 20 = 400 con-

nections). The sLORETA also computed the source

reconstruction algorithm previously described (Pascual-

Marqui and Biscay-Lirio 1993; Pascual-Marqui et al. 1994,

1995).

Statistical analysis

Power spectra analysis and EEG connectivity (lagged

coherence) were compared among conditions, for each

frequency band. The conditions analyzed were three:

baseline (BL), AM-T, and exam recall (Task 2). All

comparisons were performed by using the statistical non-

parametric mapping methodology supplied by the sLO-

RETA (Nichols and Holmes 2002). This methodology is

based on the Fisher’s permutation test: a subset of non-

parametric statistics. In particular, this is a type of statis-

tical significance test in which the distribution of the test

statistic under the null hypothesis is obtained by calculating

all possible values of the test statistic under rearrangements

of the labels on the observed data points. Correction of

significance for multiple testing was computed for the two

comparisons between conditions for each frequency band:

for the correction, we applied the nonparametric random-

ization procedure available in the sLORETA program

package (Nichols and Holmes 2002).

T-level thresholds were computed by the statistical

software implemented in the sLORETA, which correspond

to the statistical significance thresholds (p \ 0.05 and

p \ 0.01) (for details see, Friston et al. 1990; Friston et al.

1991).

Results

EEG recordings suitable for the analysis were obtained in

all cases. Visual evaluation of the EEG recordings showed

no relevant modifications of the background rhythm fre-

quency, focal abnormalities, or epileptic discharges. No

subject showed evidence of drowsiness or sleep during the

recordings. In a post-session interview, all subjects repor-

ted that they had no difficulties and no major distractions

while performing the tasks.

Despite our statistical tests were two-tailed, all our

comparisons generated positive T values.

Power spectra analysis

In the comparison between AM-T and BL, the thresholds

for significance were T = 3.554, corresponding to

p \ 0.05, and T = 4.559, corresponding to p \ 0.01. Sig-

nificant modifications were documented in the delta

(0.5–4 Hz) frequency band: in the AM-T condition,

increased power of delta activity was observed in frontal

and midline cortices. sLORETA software localized these

modification in the dorsolateral PFC (Brodmann areas, BA

9; T = 3.682, corresponding to p = 0,043), in anterior

PFC (BA 10; T = 3.801, corresponding to p = 0.038), in

orbitofrontal cortex (OFC, BA 11; T = 4.323, corre-

sponding to p = 0.016), in ventral anterior cingulate cortex

(ACC, BA 32, T = 4.204, corresponding to p = 0.021),

and in anterior (dorsal and ventral) cingulate cortex (BA

24; T = 4.086, corresponding to p = 0.026) (Fig. 2a).

In the comparison between Task 2 and BL, the thresh-

olds for significance were T = 3.960, corresponding to

p \ 0.05, and T = 4.872, corresponding to p \ 0.01. The

only significant modification was observed in the gamma

(30–60 Hz): during Task 2, an increase in gamma power

was recorded in the posterior midline areas. sLORETA

software localized this modification in left precuneus (BA

7; T = 4.169, corresponding to p = 0.025) (Fig. 2b).

Finally, in the comparison between AM-T and Task 2,

the thresholds for significance were T = 8.808,

Cogn Process

123

Fig

.2

Res

ult

so

fth

esL

OR

ET

Aco

mp

aris

on

of

EE

Gp

ow

ersp

ectr

ain

all

freq

uen

cyb

and

s.P

an

elA

AM

-Tv

ersu

sB

asel

ine;

Pa

nel

BT

ask

2v

ersu

sB

asel

ine.

Co

lore

dsp

ots

ind

icat

ear

eas

wh

ere

stat

isti

call

ysi

gn

ifica

nt

incr

ease

ind

elta

EE

Gsp

ectr

alp

ow

erw

asm

easu

red

.L

evel

so

fsi

gn

ifica

nce

are

rep

rese

nte

din

the

cen

ter

of

the

fig

ure

.R

edco

lou

rin

dic

ates

sig

nifi

can

tin

crea

sein

po

wer

spec

tra;

blu

eco

lou

rin

dic

ates

sig

nifi

can

tre

du

ctio

nin

po

wer

spec

tra.

Th

resh

old

val

ues

(T)

for

stat

isti

cal

sig

nifi

can

ce(c

orr

esp

on

din

gto

p\

0.0

5)

are

rep

ort

edin

the

fig

ure

cen

ter.

InA

M-T

,

sig

nifi

can

tm

od

ifica

tio

ns

wer

eo

bse

rved

ind

elta

ban

din

do

rso

late

ral

PF

C(B

A9

),in

ante

rio

rP

FC

(BA

10

),in

OF

C(B

A1

1),

inv

entr

al,

and

inA

CC

(BA

32

,2

4).

InT

ask

2,

sig

nifi

can

t

mo

difi

cati

on

sw

ere

ob

serv

edin

del

tab

and

inth

ele

ftp

recu

neu

s(B

A7

).B

AB

rod

man

nar

eas,

LH

left

hem

isp

her

e,A

ante

rio

r,P

po

ster

ior

Cogn Process

123

corresponding to p \ 0.05, and T = 8.821, corresponding

to p \ 0.01. No significant differences were observed, in

all frequency bands.

Lagged coherence analysis

In the comparison between AM-T and BL, the thresholds

for significance were T = 4.154, corresponding to

p \ 0.05, and T = 4,984, corresponding to p \ 0.01. Sig-

nificant modifications were observed in the delta

(0.5–4 Hz; T = 4.522, corresponding to p = 0.031) as well

as in the gamma (30–60 Hz; T = 4.346, corresponding to

p = 0.039) frequency bands. In the delta band, the AM-T

condition was associated with a widespread increase in

lagged coherence, involving the prefrontal, temporal,

parietal, and occipital ROIs (Fig. 3a). In the gamma band,

AM-T was associated with increased coherence between

mesial temporal and left parieto-occipital ROIs (Fig. 3a).

In the comparison between Task 2 and BL, the thresh-

olds for significance were T = 4.579, corresponding to

p \ 0.05, and T = 5.572, corresponding to p \ 0.01. As

for AM-T, significant modifications involved the delta

(0.5–4 Hz; T = 5.236, corresponding to p = 0.027) as well

as the gamma (30–60 Hz; T = 4.904, corresponding to

p = 0.041) bands. In the delta band, the Task 2 condition

was associated with a widespread increase in lagged

coherence, involving the prefrontal, temporal, parietal, and

occipital ROIs (Fig. 3b). In the gamma band, Task 2 was

associated with increased coherence between mesial tem-

poral and right parieto-occipital ROIs (Fig. 3b).

Finally, in the comparison between AM-T and Task 2,

the thresholds for significance were T = 4.096, corre-

sponding to p \ 0.05, and T = 4,901, corresponding to

p \ 0.01. No significant differences were observed, in all

frequency bands.

Discussion

The principal aim of this study was to explore the modi-

fications of EEG power spectra and connectivity induced

by two different AM tasks: a standardized, validated AM

task (Williams and Broadbent 1986) and a more ecological

retrieval of a single autobiographical event.

The results indicate that the AM-T is associated with

increased delta band power in left PFC and in ACC:

bilateral, widespread increase in EEG connectivity also in

the delta band, and increase in left temporo-parietal con-

nections in the gamma band.

Conversely, the Task 2 was characterized by an increase

in gamma power in left parieto-occipital cortex, and sim-

ilarly to AM-T, a widespread increase in EEG connectivity

also in the delta band. Moreover, in Task 2, we observed anFig

.3

Res

ult

so

fth

esL

OR

ET

Aco

mp

aris

on

of

EE

Gla

gg

edco

her

ence

inal

lfr

equ

ency

ban

ds.

Pa

nel

AA

M-T

ver

sus

Bas

elin

e;P

an

elB

Tas

k2

ver

sus

Bas

elin

e.T

hre

sho

ldv

alu

es(T

)fo

r

stat

isti

cal

sig

nifi

can

ce(c

orr

esp

on

din

gto

p\

0.0

5)

are

rep

ort

edin

cen

ter

of

the

fig

ure

;re

dli

nes

ind

icat

eco

nn

ecti

on

sw

hic

hp

rese

nte

dsi

gn

ifica

nt

incr

ease

inco

her

ence

;b

lue

lin

es(n

ot

pre

sen

t)

wo

uld

ind

icat

esi

gn

ifica

nt

red

uct

ion

inco

her

ence

.R

rig

ht,

Lle

ft,

Aan

teri

or,

Pp

ost

erio

r

Cogn Process

123

increase in right temporo-parietal connections in the

gamma band.

The engagement of these brain areas in AM is widely

documented (Cabeza and St Jacques 2007; Maguire 2001;

Svoboda et al. 2006). Different regions of frontal lobes

play a crucial role in AM. In AM-T, we documented the

involvement of dorsolateral PFC (BA 9), in anterior PFC

(BA 10), and in OFC (BA 11).

Medial frontal region (BA 9, 10) seems to be critical in

AM. Indeed, it is documented that this region is involved in

self-referential processing during AMs (Addis et al. 2004;

Cabeza et al. 2004; Levine et al. 2004; Macrae et al. 2004).

This process seems to be the key element of AM so that

several authors (Conway 2005; Conway and Pleydell-Pe-

arce 2000; Conway et al. 2001, 2003) proposed that auto-

biographical memory is constructed within a self-memory

system (SMS), a conceptual model which consists of two

main components: the working self and the autobiograph-

ical memory knowledge base. When these components

interlock in acts of remembering, specific autobiographical

memories can be formed.

Furthermore, the increased delta power in OFC,

observed in our study, could reflect the emotional infor-

mation processing during the AM-T: this is consistent with

previous findings which documented the engagement of

OFC in emotional AM task (Maddock et al. 2001; Mark-

owitsch et al. 2003; Piefke et al. 2003).

The AM-T, as compared with BL, also provoked an

increase in delta power in the anterior cingulate cortex.

Although the role of posterior cingulate cortex in AM has

been established by neuroimaging studies (Svoboda et al.

2006), also ACC seems to play an important role in

memory reconstruction and monitoring process (Cabeza

and Nyberg 2000; Duncan and Owen 2000; Fletcher and

Henson 2001) of AM. BA 24 and 32 are considered

responsible for the cognitive components of ACC (De-

vinsky et al. 1995), and the interaction with PFC could

reflect the response to the erroneous information during

monitoring process of AMs. According to the ‘conflict-

monitoring’ theory (Carter et al. 1999, 2000), it has been

suggested that ACC provides an online conflict signal,

indicating the need to engage brain regions such as dor-

solateral prefrontal cortex and inferior parietal cortex to

implement strategic processes (Carter et al. 1999, 2000).

This has already been documented in different cognitive

tasks such as the Stroop Task (MacDonald et al. 2000).

Our results also documented an increase in gamma band

localized in the left precuneus (BA 7) in Task 2 when

compared with BL. This structure is involved in self-ref-

erential processing, visuo-spatial imagery, and episodic

memory retrieval (Cavanna and Trimble 2006), and it

seems to be an important area in the cortical AM network

(Bullmore and Sporns 2009). In the perspective of AM,

several studies suggest the implication of precuneus in

visual imagery (Addis et al. 2004; Gardini et al. 2006). It is

possible to speculate that the greater activation of precu-

neus in this task could reflect the vividness of a very

stressful event such as the high school exit examination.

The results of the EEG connectivity analysis suggest

that the AM tasks are associated with increased cortical

connectivity, both in the low-frequency (delta) and in the

high-frequency (gamma) EEG bands. The two different

tasks used in this study showed similar effects in the delta

frequency band: in both tasks, a spread bilateral activation

of a cortical network was observed (Fig. 3a, b). Con-

versely, the two AM tasks produced different, localized

changes in the gamma band: while the AM-T was associ-

ated with increased connectivity in the left temporo-pari-

etal areas, the Task 2 induced a similar increase in

connectivity in the right hemisphere.

In this respect, the results are consistent with the find-

ings of functional imaging and neurophysiological studies

of AM described in the literature (Cabeza and St Jacques

2007; Conway et al. 2001, 2003; Maguire 2001; Steinvorth

et al. 2010; Svoboda et al. 2006). In both AM tasks, we

observed two similar EEG connectivity patterns: the first is

associated with the delta band, and it corresponds to a wide

and complex brain network, involving bilateral prefrontal,

temporal, and occipital areas; the second is associated with

the gamma band, and it corresponds to smaller brain net-

works involving posterior areas.

Bilateral activation of functional connections observed

in the first network could reflect the emotional valence of

our AM tasks. The predominant activation of left-side areas

reported in previous studies (Cabeza and St Jacques 2007;

Maguire 2001; Svoboda et al. 2006) is thought to reflect the

contribution of semantic information to AM neutral

retrieval cues (Cabeza and St Jacques 2007). Nevertheless,

AM is also characterized by emotional content and vivid

sensory details (Rubin 2006), and several researches doc-

umented right-lateralized or bilateral activation patterns

during emotional AM tasks (Denkova et al. 2006; Fink

et al. 1996; Markowitsch et al. 2000; Vandekerckhove

et al. 2005). Our results are consistent with Vandekerckh-

ove et al. (2005) which reported bi-hemispheric activation

(including MTL and PCF) during stressful, negative or

positive, AM tasks. Furthermore, we showed that this

complex network is supported by the delta frequency band.

This is consistent with different studies, reporting the

involvement of delta frequency in memory. It is proposed

that delta oscillation plays a crucial role in memory con-

solidation during sleep (Sagar et al. 1985; Sirota et al.

2003), successful explicit memory formation (Fell et al.

2006), working memory (Harmony et al. 1996; Imperatori

et al. 2013), and with AM (Steinvorth et al. 2010). More-

over, the involvement of delta band could reflect the

Cogn Process

123

neurophysiological index of the inter-neuronal exchange

along the distributed brain areas that shape the AM cortical

network (Steinvorth et al. 2010; Toth et al. 2012).

Our results also documented an increase in EEG

coherence in gamma band during both AM tasks in pos-

terior areas. Whereas AM-T is characterized by increase in

gamma coherence in the left hemisphere, Task 2 is char-

acterized by increase in gamma coherence in the right

hemisphere (Fig. 3). The difference between left and right

connectivity could reflect the difference among the two

AM Tasks: in AM-T, participants were instructed to recall

different AMs in response to verbally presented cue words,

whereas in Task 2 we only asked to remember and keep in

mind details about a single autobiographical event.

The crucial role of gamma bands in memory is widely

documented (Duzel et al. 2010; Jutras and Buffalo 2010). In

the present research, the increase in gamma coherence

between parietal and temporal areas could reflect the use of

information arising from correspondences between the cues

and material from the long-term memory. This is in line with

the results reported by Steinvorth et al. (2010), who observed

an increase in gamma power in superficial layers of en-

thorhinal cortex beginning at 200 ms after the cue and lasting

until 12,000 ms at the end of the analyzed time period.

It must be specified that we did not document modi-

fications of EEG coherence and EEG power spectra in the

theta band. In a study by Corsi-Cabrera et al. (2000),

power spectra from wake and sleep in healthy adults were

submitted to principal component analyses to investigate

which frequencies covaried together (Corsi-Cabrera et al.

2000). The results indicated that slow-wave activity can

oscillate at higher frequencies, up to 8 Hz; interestingly,

no theta band was independently identified: According to

the authors, this suggested either that delta and theta

oscillations are two rhythms under the same global

influence or that the traditional division of theta band in

the human cortical EEG is artificial (Corsi-Cabrera et al.

2000).

Furthermore, Lega et al. (2012) recording intracranial

EEG during an EM task revealed that only ‘slow-theta’

(2.5–5 Hz) oscillation was functionally linked to gamma

oscillations suggesting that this EEG pattern plays an

important role in cortico–hippocampal communication.

Moreover, we did not find any significant modification

in the beta bands, although this activity seems to play an

important role in specific memory tasks such as visuo-

spatial working memory (Roberts et al. 2013). However,

the role of beta activity in AM remains controversial and

not constantly documented (Conway et al. 2001, 2003;

Steinvorth et al. 2010).

In conclusion, our findings indicate that AM retrieval

engages in a complex network which is mediated by both

low- (delta) and high-frequency (gamma) EEG bands.

Study limitations

The present study has main limitations. The first is the use

of scalp EEG recordings, which have an intrinsic limit in

space resolution, particularly in the identification of deep

subcortical sources. A further limit is in the montage

applied, which is the one used in standard EEG recording.

It is known that spatial resolution of EEG sources increases

with the number of electrodes, and therefore, high-density

recordings are more reliable in the esteem of EEG rhythms

source analysis. The same kind of limitation, obviously, is

reflected by the sLORETA software, which is, by defini-

tion, a low-resolution electric source analysis software.

Furthermore, it is possible that the modification observed in

the gamma band was influenced by the superimposed

cranial and ocular muscles artifacts, which is particularly

evident in gamma activity (Hipp and Siegel 2013). Finally,

although it is widely used, we compared the AM tasks with

a resting state and this needs caution in the interpretation of

results (Svoboda et al. 2006).

Conflict of interest None.

References

Addis DR, Moscovitch M, Crawley AP, McAndrews MP (2004)

Recollective qualities modulate hippocampal activation during

autobiographical memory retrieval. Hippocampus 14(6):752–762.

doi:10.1002/hipo.10215

Bullmore E, Sporns O (2009) Complex brain networks: graph

theoretical analysis of structural and functional systems. Nat Rev

Neurosci 10(3):186–198. doi:10.1038/nrn2575

Cabeza R, Nyberg L (2000) Imaging cognition II: an empirical review

of 275 PET and fMRI studies. J Cogn Neurosci 12(1):1–47.

doi:10.1162/08989290051137585

Cabeza R, St Jacques P (2007) Functional neuroimaging of autobio-

graphical memory. Trends Cogn Sci 11(5):219–227. doi:10.

1016/j.tics.2007.02.005

Cabeza R, Prince SE, Daselaar SM, Greenberg DL, Budde M, Dolcos

F, LaBar KS, Rubin DC (2004) Brain activity during episodic

retrieval of autobiographical and laboratory events: an fMRI

study using a novel photo paradigm. J Cogn Neurosci

16(9):1583–1594. doi:10.1162/0898929042568578

Carter CS, Botvinick MM, Cohen JD (1999) The contribution of the

anterior cingulate cortex to executive processes in cognition. Rev

Neurosci 10(1):49–57

Carter CS, Macdonald AM, Botvinick M, Ross LL, Stenger VA, Noll

D, Cohen JD (2000) Parsing executive processes: strategic vs.

evaluative functions of the anterior cingulate cortex. Proc Natl

Acad Sci USA 97(4):1944–1948. doi:10.1073/panas.97/4/1944

Cavanna AE, Trimble MR (2006) The precuneus: a review of its

functional anatomy and behavioural correlates. Brain 129(Pt

3):564–583. doi:10.1093/brain/awl004

Conway MA (2005) Memory and the self. J Mem Lang

53(4):594–628. doi:10.1016/j.jml.2005.08.005

Conway MA, Pleydell-Pearce CW (2000) The construction of

autobiographical memories in the self-memory system. Psychol

Rev 107(2):261–288. doi:10.1037//0033-295X

Conway MA, Pleydell-Pearce CW, Whitecross SE (2001) The

neuroanatomy of autobiographical memory: a slow cortical

Cogn Process

123

potential study of autobiographical memory retrieval. J Mem

Lang 45:493–524. doi:10.1006/jmla.2001.2781

Conway MA, Pleydell-Pearce CW, Whitecross SE, Sharpe H (2003)

Neurophysiological correlates of memory for experienced and

imagined events. Neuropsychologia 41(3):334–340. doi:10.

1016/S0028393202001653

Corsi-Cabrera M, Guevara MA, Del Rio-Portilla Y, Arce C,

Villanueva-Hernandez Y (2000) EEG bands during wakefulness,

slow-wave and paradoxical sleep as a result of principal

component analysis in man. Sleep 23(6):738–744

Daselaar SM, Rice HJ, Greenberg DL, Cabeza R, LaBar KS, Rubin DC

(2008) The spatiotemporal dynamics of autobiographical mem-

ory: neural correlates of recall, emotional intensity, and reliving.

Cereb Cortex 18(1):217–229. doi:10.1093/cercor/bhm048

Denkova E, Botzung A, Scheiber C, Manning L (2006) Implicit emotion

during recollection of past events: a nonverbal fMRI study. Brain

Res 1078(1):143–150. doi:10.1016/j.brainres.2006.01.061

Devinsky O, Morrell MJ, Vogt BA (1995) Contributions of anterior

cingulate cortex to behaviour. Brain 118(Pt 1):279–306

Duncan J, Owen AM (2000) Common regions of the human frontal

lobe recruited by diverse cognitive demands. Trends Neurosci

23(10):475–483. doi:10.1016/S0166-2236(00)01633-7

Duzel E, Penny WD, Burgess N (2010) Brain oscillations and

memory. Curr Opin Neurobiol 20(2):143–149. doi:10.1016/j.

conb.2010.01.004

Farina B, Speranza AM, Dittoni S, Gnoni V, Trentini C, Vergano CM,

Liotti G, Brunetti R, Testani E, Marca GD (2013) Memories of

attachment hamper EEG cortical connectivity in dissociative

patients. Eur Arch Psychiatry Clin Neurosci. doi:10.1007/

s00406-013-0461-9

Fell J, Fernandez G, Klaver P, Axmacher N, Mormann F, Haupt S,

Elger CE (2006) Rhinal-hippocampal coupling during declara-

tive memory formation: dependence on item characteristics.

Neurosci Lett 407(1):37–41. doi:10.1016/j.neulet.2006.07.074

Fink GR, Markowitsch HJ, Reinkemeier M, Bruckbauer T, Kessler J,

Heiss WD (1996) Cerebral representation of one’s own past:

neural networks involved in autobiographical memory. J Neuro-

sci 16(13):4275–4282

Fletcher PC, Henson RN (2001) Frontal lobes and human memory:

insights from functional neuroimaging. Brain 124(Pt

5):849–881. doi:10.1093/brain/124.5.849

Friston KJ, Frith CD, Liddle PF, Dolan RJ, Lammertsma AA,

Frackowiak RS (1990) The relationship between global and local

changes in PET scans. J Cereb Blood Flow Metab

10(4):458–466. doi:10.1038/jcbfm.1990.88

Friston KJ, Frith CD, Liddle PF, Frackowiak RS (1991) Comparing

functional (PET) images: the assessment of significant change.

J Cereb Blood Flow Metab 11(4):690–699. doi:10.1038/jcbfm.

1991.122

Gardini S, Cornoldi C, De Beni R, Venneri A (2006) Left

mediotemporal structures mediate the retrieval of episodic

autobiographical mental images. Neuroimage 30(2):645–655.

doi:10.1016/j.neuroimage.2005.10.012

Grave de Peralta Menendez R, Gonzalez Andino SL, Morand S,

Michel CM, Landis T (2000) Imaging the electrical activity of

the brain: ELECTRA. Hum Brain Mapp 9(1):1–12

Grave de Peralta-Menendez R, Gonzalez-Andino SL (1998) A critical

analysis of linear inverse solutions to the neuroelectromagnetic

inverse problem. IEEE Trans Biomed Eng 45(4):440–448.

doi:10.1109/10.664200

Harmony T, Fernandez T, Silva J, Bernal J, Diaz-Comas L, Reyes A,

Marosi E, Rodriguez M (1996) EEG delta activity: an indicator

of attention to internal processing during performance of mental

tasks. Int J Psychophysiol 24(1–2):161–171

Hipp JF, Siegel M (2013) Dissociating neuronal gamma-band activity

from cranial and ocular muscle activity in EEG. Front Hum

Neurosci 7:338. doi:10.3389/fnhum.2013.00338

Holland AC, Addis DR, Kensinger EA (2011) The neural correlates

of specific versus general autobiographical memory construction

and elaboration. Neuropsychologia 49(12):3164–3177. doi:10.

1016/j.neuropsychologia.2011.07.015

Imperatori C, Farina B, Brunetti R, Gnoni V, Testani E, Quintiliani

MI, Del Gatto C, Indraccolo A, Contardi A, Speranza AM, Della

Marca G (2013) Modifications of EEG power spectra in mesial

temporal lobe during n-back tasks of increasing difficulty. A

sLORETA study. Front Hum Neurosci 7:109. doi:10.3389/

fnhum.2013.00109

Jutras MJ, Buffalo EA (2010) Synchronous neural activity and

memory formation. Curr Opin Neurobiol 20(2):150–155. doi:10.

1016/j.conb.2010.02.006

Kreiter AK, Singer W (1992) Oscillatory neuronal responses in the

visual cortex of the awake macaque monkey. Eur J Neurosci

4(4):369–375

Lega BC, Jacobs J, Kahana M (2012) Human hippocampal theta

oscillations and the formation of episodic memories. Hippocam-

pus 22(4):748–761. doi:10.1002/hipo.20937

Levine B, Turner GR, Tisserand D, Hevenor SJ, Graham SJ,

McIntosh AR (2004) The functional neuroanatomy of episodic

and semantic autobiographical remembering: a prospective

functional MRI study. J Cogn Neurosci 16(9):1633–1646.

doi:10.1162/0898929042568587

MacDonald AW 3rd, Cohen JD, Stenger VA, Carter CS (2000)

Dissociating the role of the dorsolateral prefrontal and anterior

cingulate cortex in cognitive control. Science 288(5472):

1835–1838. doi:10.1126/science.288.5472.1835

Macrae CN, Moran JM, Heatherton TF, Banfield JF, Kelley WM

(2004) Medial prefrontal activity predicts memory for self.

Cereb Cortex 14(6):647–654. doi:10.1093/cercor/bhh025

Maddock RJ, Garrett AS, Buonocore MH (2001) Remembering

familiar people: the posterior cingulate cortex and autobiograph-

ical memory retrieval. Neuroscience 104(3):667–676. doi:10.

1016/S0306-4522(01)00108-7

Maguire EA (2001) Neuroimaging studies of autobiographical event

memory. Philos Trans R Soc Lond B Biol Sci 356(1413):

1441–1451. doi:10.1098/rstb.2001.0944

Markowitsch HJ, Thiel A, Reinkemeier M, Kessler J, Koyuncu A,

Heiss WD (2000) Right amygdalar and temporofrontal activation

during autobiographic, but not during fictitious memory

retrieval. Behav Neurol 12(4):181–190

Markowitsch HJ, Vandekerckhove MM, Lanfermann H, Russ MO

(2003) Engagement of lateral and medial prefrontal areas in the

ecphory of sad and happy autobiographical memories. Cortex

39(4–5):643–665. doi:10.1016/S0010-9452(08)70858-X

Moradi AR, Herlihy J, Yasseri G, Shahraray M, Turner S, Dalgleish T

(2008) Specificity of episodic and semantic aspects of autobio-

graphical memory in relation to symptoms of posttraumatic

stress disorder (PTSD). Acta Psychol (Amst) 127(3):645–653.

doi:10.1016/j.actpsy.2007.11.001

Murphy TH, Blatter LA, Wier WG, Baraban JM (1992) Spontaneous

synchronous synaptic calcium transients in cultured cortical

neurons. J Neurosci 12(12):4834–4845

Nichols TE, Holmes AP (2002) Nonparametric permutation tests for

functional neuroimaging: a primer with examples. Hum Brain

Mapp 15(1):1–25. doi:10.1002/hbm.1058

Nolte G, Bai O, Wheaton L, Mari Z, Vorbach S, Hallett M (2004)

Identifying true brain interaction from EEG data using the

imaginary part of coherency. Clin Neurophysiol 115(10):

2292–2307. doi:10.1016/j.clinph.2004.04.029

Cogn Process

123

Pascual-Marqui RD (2007a) Coherence and phase synchronization:

generalization to pairs of multivariate time series, and removal

of zero-lag contributions. arXiv:07061776v3 [statME] 12 July

2007 (http://www.arxivorg/pdf/07061776)

Pascual-Marqui RD (2007b) Instantaneous and lagged measurements

of linear and nonlinear dependence between groups of multivariate

time series: frequency decomposition. arXiv:07111455 [statME],

2007-November-09, http://www.arxivorg/abs/07111455

Pascual-Marqui RD, Biscay-Lirio R (1993) Spatial resolution of

neuronal generators based on EEG and MEG measurements. Int

J Neurosci 68(1–2):93–105

Pascual-Marqui RD, Michel CM, Lehmann D (1994) Low resolution

electromagnetic tomography: a new method for localizing

electrical activity in the brain. Int J Psychophysiol

18(1):49–65. doi:10.1016/0167-8760(84)90014-X

Pascual-Marqui RD, Michel CM, Lehmann D (1995) Segmentation of

brain electrical activity into microstates: model estimation and

validation. IEEE Trans Biomed Eng 42(7):658–665. doi:10.

1109/10.391164

Pascual-Marqui RD, Lehmann D, Koukkou M, Kochi K, Anderer P,

Saletu B, Tanaka H, Hirata K, John ER, Prichep L, Biscay-Lirio

R, Kinoshita T (2011) Assessing interactions in the brain with

exact low-resolution electromagnetic tomography. Philos Trans

A Math Phys Eng Sci 369(1952):3768–3784. doi:10.1098/rsta.

2011.0081

Piefke M, Weiss PH, Zilles K, Markowitsch HJ, Fink GR (2003)

Differential remoteness and emotional tone modulate the neural

correlates of autobiographical memory. Brain 126(Pt

3):650–668. doi:10.1093/brain/awg064

Roberts BM, Hsieh LT, Ranganath C (2013) Oscillatory activity

during maintenance of spatial and temporal information in

working memory. Neuropsychologia 51(2):349–357. doi:10.

1016/j.neuropsychologia

Rubin DC (2006) The basic-systems model of episodic memory.

Perspect Psychol Sci 1(4):277–311. doi:10.1111/j.1745-6916.

2006.00017X

Rugg MD, Vilberg KL (2013) Brain networks underlying episodic

memory retrieval. Curr Opin Neurobiol 23(2):255–260. doi:10.

1016/j.conb.2012.11.005

Sagar HJ, Cohen NJ, Corkin S, Growdon JH (1985) Dissociations

among processes in remote memory. Ann N Y Acad Sci

444:533–535

Sirota A, Csicsvari J, Buhl D, Buzsaki G (2003) Communication

between neocortex and hippocampus during sleep in rodents.

Proc Natl Acad Sci USA 100(4):2065–2069. doi:10.1073/pnas.

0437938100

Steinvorth S, Wang C, Ulbert I, Schomer D, Halgren E (2010) Human

entorhinal gamma and theta oscillations selective for remote

autobiographical memory. Hippocampus 20(1):166–173. doi:10.

1002/hipo.20597

Svoboda E, McKinnon MC, Levine B (2006) The functional

neuroanatomy of autobiographical memory: a meta-analysis.

Neuropsychologia 44(12):2189–2208. doi:10.1016/j.neuropsy

chologia.2006.05.023

Toth B, Boha R, Posfai M, Gaal ZA, Konya A, Stam CJ, Molnar M

(2012) EEG synchronization characteristics of functional con-

nectivity and complex network properties of memory mainte-

nance in the delta and theta frequency bands. Int J Psychophysiol

83(3):399–402. doi:10.1016/j.ijpsycho.2011.11.017

Tulving E (1985) Memory and consciousness. Can Psychol

26(1):1–12

Tulving E (1987) Multiple memory systems and consciousness. Hum

Neurobiol 6(2):67–80

Van Vreeswijk MF, De Wilde EJ (2004) Autobiographical memory

specificity, psychopathology, depressed mood and the use of the

autobiographical memory test: a meta-analysis. Behav Res Ther

42(6):731–743. doi:10.1016/S0005-7967(03)00194-3

Vandekerckhove MM, Markowitsch HJ, Mertens M, Woermann FG

(2005) Bi-hemispheric engagement in the retrieval of autobio-

graphical episodes. Behav Neurol 16(4):203–210

Williams JM, Broadbent K (1986) Autobiographical memory in

suicide attempters. J Abnorm Psychol 95(2):144–149

Williams JM, Barnhofer T, Crane C, Herman D, Raes F, Watkins E,

Dalgleish T (2007) Autobiographical memory specificity and

emotional disorder. Psychol Bull 133(1):122–148. doi:10.1037/

0033-2909.133.1.122

Zijlmans M, Huiskamp GM, van Huffelen AC, Spetgens WP, Leijten

FS (2008) Detection of temporal lobe spikes: comparing

nasopharyngeal, cheek and anterior temporal electrodes to

simultaneous subdural recordings. Clin Neurophysiol

119(8):1771–1777. doi:10.1016/j.clinph.2008.04.011

Cogn Process

123