Spatiotemporal pattern of brain electrical activity related to immediate and delayed episodic memory retrieval

Contents lists available at ScienceDirect Neurobiology of Learning and Memory journal homepage: www.elsevier.com/locate/ynlme Spatiotemporal pattern of brain electrical activity related to immediate and delayed episodic memory retrieval Miguel Ángel Rivas-Fernández ⁎ , Santiago Galdo-Álvarez, Montserrat Zurrón, Fernando Díaz, Mónica Lindín Laboratorio de Neurociencia Cognitiva, Departamento de Psicoloxía Clínica e Psicobioloxía, Universidade de Santiago de Compostela, Galicia, Spain ARTICLE INFO Keywords: Episodic memory Recognition memory Event-related brain potentials (ERPs) Old/new ERP effects eLORETA ABSTRACT In the present study we used the event-related brain potentials (ERP) technique and eLORETA (exact lowresolution electromagnetic tomography) method in order to characterize and compare the performance and the spatiotemporal pattern of the brain electrical activity related to the immediate episodic retrieval of information (words) that is being learned relative to delayed episodic retrieval twenty-minutes later. For this purpose, 16 young participants carried out an old/new word recognition task with source memory (word colour). The task included an immediate memory phase (with three study-test blocks) followed (20 min later) by a delayed memory phase with one test block. The behavioural data showed progressive learning and consolidation of the information (old words) during the immediate memory phase. The ERP data to correctly identified old words for which the colour was subsequently recollected (H/H) compared to the correctly rejected new words (CR) showed: (1) a significant more positive-going potential in the 500–675 ms post-stimulus interval (parietal old/ new effect, related to recollection), and (2) a more negative-going potential in the 950–1850 ms interval (LPN effect, related to retrieval and post-retrieval processes). The eLORETA data also revealed that the successful recognition of old words (and probably retrieval of their colour) was accompanied by activation of (1) left medial temporal (parahippocampal gyrus) and parietal regions involved in the recollection in both memory phases, and (2) prefrontal regions and the superior temporal gyrus (in the immediate and delayed memory phases respectively) involved in monitoring, evaluating and maintaining the retrieval products. These findings indicate that episodic memory retrieval depends on a network involving medial temporal lobe and frontal, parietal and temporal neocortical structures. That network was involved in immediate and delayed memory retrieval and during the course of memory consolidation, with greater activation of some nodes (mobilization of more processing resources) for the delayed respect to the immediate retrieval condition. 1. Introduction Episodic memory (EM), defined as a neurocognitive system that enables us to consciously recall past experiences (Tulving, 2002), has received increasing attention in cognitive, neuropsychological, psychophysiological and neuroimaging studies in the past decade (Rugg & Vilberg, 2013; Rugg & Yonelinas, 2003; Yonelinas, 2002). EM studies have provided insights into how information is acquired, organized and retrieved in long-term memory, by evaluating (1) encoding processes (or storage of new information involving changes in the strength and/or number of synaptic connections within the nervous system, known as memory traces), (2) consolidation (stabilization of memory traces following the initial encoding), and (3) retrieval (i.e. recall of previously stored information) (Squire, Wixted, & Clark, 2007). The Standard Consolidation Theory (SCT) (Squire & Alvarez, 1995) proposed that the encoding, consolidation and retrieval of memories are always dependent on direct connections between the hippocampal formation and neocortical regions; however, once consolidation stabilizes the memory traces, the retrieval of episodic memories is supported by activation of neocortical regions that are more independent of hippocampal formation, i.e. connections between hippocampal formation and neocortical regions become less critical for retrieving the stored information. However, this view was later questioned, and it is becoming more widely accepted that most memories never become https://doi.org/10.1016/j.nlm.2020.107309 Received 6 May 2020; Received in revised form 30 July 2020; Accepted 26 August 2020 ⁎ Corresponding author at: Laboratorio de Neurociencia Cognitiva, Departamento de Psicoloxía Clínica e Psicobioloxía, Universidade de Santiago de Compostela, Campus Vida, Calle Xosé María Suárez Núñez, s/n, 15782 Santiago de Compostela, Galicia, Spain. E-mail addresses: [email protected] (M.Á. Rivas-Fernández), [email protected] (S. Galdo-Álvarez), [email protected] (M. Zurrón), [email protected] (F. Díaz), [email protected] (M. Lindín). Neurobiology of Learning and Memory 175 (2020) 107309 Available online 02 September 2020 1074-7427/ © 2020 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). T independent of hippocampal formation (see Skelin, Kilianski, & McNaughton, 2019). The Multiple Trace Theory (MTT) (Nadel & Moscovitch, 1997) and the later Trace Transformation Theory (TTT) (Sekeres, Winocur, & Moscovitch, 2018) postulate that the repeated activation of memories creates new memory traces and that the retrieval of episodic memories (recent or remote) will always depend on activation of hippocampal formation. The three aforementioned theories were proposed in relation to the systems consolidation. Two essential components of the consolidation process were proposed (Dudai, 2004; Dudai, Karni, & Born, 2015; Sekeres et al., 2018; Winocur & Moscovitch, 2011): (1) a cellular/synaptic component (synaptic consolidation, which is achieved within minutes to hours), related to early and rapid changes at local synaptic connections and cellular nodes in local neuronal assemblies. It is mediated by intracellular molecular mechanisms, and refers to the postencoding transformation of information into a long-term form; and (2) a systems component (systems consolidation, with variation in its duration, from minutes to decades), which is associated with the post-encoding time-dependent reorganization of long-term memory representations over distributed (medial temporal-neocortical) networks. Both were considered as a part of a continuous and dynamic process, so that synaptic consolidation could be deemed as subroutines into systems consolidation process (Dudai et al., 2015). Neuroimaging and behavioural studies in humans have demonstrated the importance of the medial temporal lobe (MTL) in episodic memory (Eichenbaum, Yonelinas, & Ranganath, 2007; Nadel & Moscovitch, 1997; Squire et al., 2007; Tulving, 2002). EM processes imply the continuous exchange of information in a network of brain areas centered on the medial temporal lobe (MTL) and neocortical regions (Battaglia, Benchenane, Sirota, Pennartz, & Wiener, 2011; Ison, Quian Quiroga, & Fried, 2015). However, as far as we are aware, no previous studies have used a brain activity recording technique with high temporal resolution to disentangle the role of the MTL and neocortical regions in the successful episodic retrieval at different times across memory consolidation. With the aim of addressing this lack of information, in the present study we applied exact low resolution tomography (eLORETA) software to event-related brain potential (ERP) data, obtained during two different conditions of successful EM retrieval: immediate and delayed (twenty minutes later). The ERP technique, which has a high temporal resolution, has been widely used in psychophysiological studies to evaluate the time course of brain electrical activity during different EM processes. Recognition memory is commonly evaluated using old/new recognition tasks, in which participants must classify items as previously presented (old stimuli) or not presented (new stimuli, correct rejection). The difference between the ERPs elicited by correctly rejected new items (CR) and the ERPs elicited by correctly recognized old items (Hit) reflects EM retrieval phenomena (Friedman, 2013). Specifically, between 300 and 500 ms post-stimulus, a negative ERP component is observed at frontal and central scalp sites, with a smaller negative voltage in response to old stimuli than in response to new stimuli: this difference is commonly known as the mid-frontal old/new effect or frontal N400 effect -FN400- (Curran, 2000). Most evidence suggests that this effect is a correlate of familiarity, as variations in amplitude are observed depending on the familiarity confidence (Woodruff, Hayama, & Rugg, 2006; Yu & Rugg, 2010). Familiarity-based recognition is considered a fast-acting, relatively automatic process that does not provide qualitative information about the study episode (Rugg & Curran, 2007). A parietal positive component is identified between 500 and 800 ms post-stimulus and with larger voltage in response to old stimuli than in response to new stimuli: this difference, which is commonly known as the parietal old/new effect (Wilding, Fraser, & Herron, 2005), is frequently larger at left scalp sites when the stimuli are words. This component has been associated with recollection, which is considered a slower and more effortful process than familiarity-based recognition and which provides information about qualitative aspects of a prior event, including its context (Rugg & Vilberg, 2013). The parietal old/ new effect shows amplitude modulations in recognition memory tasks that demand conscious recall of previously studied contextual details (source or context memory task). In particular, this effect has been found to have a larger amplitude in episodic relative to semantic (knowing facts) judgments (Macleod & Donaldson, 2017), in correct relative to incorrect source memory judgments (Macleod & Donaldson, 2017; Wilding & Rugg, 1996), and also when retrieval of a large amount of information is required (Vilberg, Moosavi, & Rugg, 2006). However, some evidence suggests that variations in the magnitude of the parietal old/new effect might not always predict variations in episodic recollection between participants (Macleod & Donaldson, 2017). Another two late onset old/new ERP effects have been identified from 600 ms onwards: (1) the right frontal old/new effect (RFE), whereby old stimuli elicit a larger positive voltage than new stimulus over right frontal scalp sites. Evidence suggests that the RFE indicates monitoring and evaluation processes that act on the retrieval products (Cruse & Wilding, 2009); and (2) the late posterior negativity (LPN) effect, which shows the opposite pattern over parieto-occipital scalp sites, i.e. old stimuli elicit a more negative-going voltage than new stimuli. Evidence suggests that LPN may reflect mnemonic processes involved in the reconstruction of a previous study episode when some item attribute is not readily retrieved or when the information retrieved needs additional evaluation, and/or non-mnemonic processes of monitoring that act in highly demanding tasks (Sommer, Vita, & De Pascalis, 2018; for a review see Mecklinger, Rosburg, & Johansson, 2016). In the present ERP study, we evaluated the old/new ERP effects in an old/new word recognition task with source memory (i.e. the colour of the old word) to characterize the spatiotemporal pattern of the brain electrical activity during the immediate and the delayed EM retrieval. Some previous studies have evaluated the old/new ERP effects in young adults, also using an old/new recognition task with words, with different temporal intervals from the study task to the recognition test: immediately (Wang, de Chastelaine, Minton, & Rugg, 2012), one minute (see experiment one of Macleod & Donaldson, 2017), 39 min (Wolk et al., 2006), one week (see experiment two of Macleod & Donaldson, 2017) and one day (Wolk et al., 2006). However, as far as we are aware, only one study (Wolk et al., 2006) have evaluated within the same experiment the old/new ERP effects using different study-test delay intervals: 39 min versus one day. In the aforementioned study, the researchers did not observe any differences in the ERP parameters (obtained in response to correctly recognized old or new words) between both delay conditions; moreover, although they identified the mid-frontal and parietal old/new effects in each delay condition, they did not evaluate the brain areas involved in these effects. Although the spatial resolution of ERPs is lower than in the functional magnetic resonance imaging (fMRI) technique, two studies have previously evaluated the spatiotemporal pattern of the brain activity related to the old/new effects (Alhaj, Massey, & McAllister-Williams, 2006; Kim et al., 2009), by applying the LORETA brain source estimation algorithm (Pascual-Marqui, Michel, & Lehmann, 1994, 2011) to ERP data. In both studies, participants performed an old/new recognition task with source memory, in which words were used as stimuli and the gender of the voice that spoke the words as the source to remember. These studies demonstrated the involvement of multiple brain areas in the successful immediate retrieval, at around 600–700 ms post-stimulus, specifically in the hippocampus, the right inferior parietal lobe, and also some regions of the superior temporal and frontal cortex. EM studies using fMRI, a technique with a much lower temporal resolution than ERP, showed high consistency in the same activated brain areas identified in the aforementioned studies in which LORETA was applied to ERP data (Bergström et al., 2013; Eichenbaum et al., 2007; Henson, Rugg, Shallice, Josephs, & Dolan, 1999; Rugg et al., 2002; Scalici, Caltagirone, & Carlesimo, 2017; Vilberg & Rugg, 2008). In summary, evaluation of old/new ERP effects allows us to assess M.Á. Rivas-Fernández, et al. Neurobiology of Learning and Memory 175 (2020) 107309 2 changes in the brain electrical activity related to immediate and delayed EM retrieval, as well as to delimit, with a good temporal resolution, the temporal ranges (considering the latency of these effects) within which we can estimate the network of different brain areas involved. Moreover, we considered that the neural data obtained (with good temporal resolution) might shed some light on the role of MTL and neocortical regions in different times of memory consolidation and retrieval processes. In the present study, we used an old/new word recognition task with source memory (word colour) to compare the performance and the spatiotemporal patterns of brain electrical activity during successful EM retrieval in an immediate memory phase (with three study-test blocks) and in a delayed memory phase (with a test presented twenty minutes later). Previous studies showed that the presentation of multiple studytest blocks (as in the immediate memory phase of the present study) improved the old/new discrimination ability of young adults when they had to recognize symbols (De Chastelaine, Friedman, Cycowicz, & Horton, 2009) and words (Jacoby, Jones, & Dolan, 1998). This improvement was reflected not only by an increase of the old/new discrimination accuracy across test repetitions (De Chastelaine et al., 2009; Jacoby et al., 1998) but also by a reduction in the reaction times during the recognition of old and new items (De Chastelaine et al., 2009). The specific aims of the study were as follows: (1) To compare performance between the immediate memory phase and the delayed memory phase. (2) To identify and evaluate the old/new ERP effects in each memory phase (immediate and delayed) and to compare these between both phases. (3) To apply eLORETA to the intervals of the old/new ERP effect identified in order to determine: a. Which brain areas are more activated during the successful EM retrieval of old words relative to the correct identification of new words, in each memory phase (immediate and delayed). b. Which brain areas are activated differently between both memory phases, during the successful retrieval of old words and during the correct identification of new words. c. The involvement of MTL and neocortical regions (e.g. prefrontal cortex and posterior parietal cortex) during the successful retrieval of old words, in each memory phase. In line with previous studies that revealed an increase in the old/ new discrimination ability in young adults during the presentation of multiple study-test blocks (De Chastelaine et al., 2009; Jacoby et al., 1998), we expected to observe a learning process of the old words across the three study/test blocks of the immediate memory phase, demonstrated behaviourally by (1) a gradual improvement in the performance across the blocks, and (2) no differences between the third block of the immediate memory phase and the test of the delayed memory phase, which would indicate that early consolidation is maintained during the 20-minute delay interval (during which participants had to perform a visuospatial working memory task: see Task in Method section). On the other hand, we expected to find the following in both memory phases: (1) old/new ERP effects, especially in the temporal ranges of the mid-frontal old/new effect related to the familiarity and the parietal old/new effect related to the recollection; and (2) greater activation of the MTL and neocortical regions during the retrieval of old words relative to the correct identification of new words. 2. Materials and method 2.1. Participants Sixteen university students (8 women, 8 men) between 18 and 25 years old (mean age: 21 years old, SD: 2.1) participated voluntarily in the study. All participants had normal audition and normal or corrected to normal vision, and none had any history of neurological or psychiatric disorders. All participants were right-handed, as evaluated by the Edinburgh Handedness Inventory (Oldfield, 1971), and all of them gave their written informed consent prior to participation in the study. The research project was approved by the Galician Clinical Research Ethics Committee (Xunta de Galicia, Spain) and was performed in accordance with the ethical standards established in the 1964 Declaration of Helsinki (Lynöe, Sandlund, Dahlqvist, & Jacobsson, 1991). 2.2. Task and stimuli 2.2.1. Task During the electroencephalographic (EEG) recording, participants performed an old/new word recognition task with source memory (word colour) and formed by an immediate memory phase and a delayed memory phase (see Fig. 1). A practice block with five words was presented before the task, to ensure that participants had understood and Fig. 1. Task scheme. M.Á. Rivas-Fernández, et al. Neurobiology of Learning and Memory 175 (2020) 107309 3 performed the task correctly. The immediate memory phase consisted of three study-test blocks, with an inter-block rest interval of 2–3 min. In the study task, the same list of 40 words (names of living beings and common objects), printed in red or blue, were randomly presented across blocks. Participants were instructed to memorize the word and the colour in which it was printed and to simultaneously make a judgment about whether the words represented living/non-living things and to respond by pressing as quickly and accurately as possible one of two different buttons on a CEDRUS Model RB-834 device (left button: index finger; right button: middle finger). The test began immediately after the study task. During each test, 26 or 27 words that had been presented in the study task (old words) were randomly interspersed with 26 or 27 new words, both printed in black (Block 1: 26 old/26 new; Block 2: 26 old/26 new; Block 3: 27 old/27 new. Thus, a total of 79 new words and 79 old words (the 40 old words presented in the study task were repeated twice, except for one word) were presented in all three tests. Participants were instructed to decide whether the word was old or new and to respond by pressing (as quickly and accurately as possible) one of two buttons (left button: index finger; right button: middle finger), depending on the category of the word: “old” if it had been presented in the study task or “new” if it had not. If the word was correctly recognized as old, participants were required to respond 500 ms later to the question “was it printed in blue?” or “was it printed in red?”, by pressing the same buttons as before (yes/no). After finishing the third block of the immediate memory phase, participants performed a different task (visuospatial working memory task, lasting 20 min, with non-verbal visual stimuli). After a short rest interval, an unexpected delayed memory test (delayed memory phase) was presented. In this phase, the 40 old words (memorized during the study task of the three blocks of the immediate memory phase) were randomly interspersed with 40 new words. All words were printed in black and participants were again instructed to press one of two different response buttons corresponding to old/new and, for correctly recognized old words, to answer the question about the colour in which words were printed in the study tasks (red or blue). Response buttons (yes/no) during the immediate and delayed memory phases were counterbalanced among participants. 2.2.2. Stimuli A total of 159 words were selected from two Spanish word databases: APO –“Animal-Persona-Objeto”- (Ferré, Guasch, Moldovan, & Sánchez-Casas, 2012) and EsPal (Duchon, Perea, Sebastián-Gallés, Martí, & Carreiras, 2013). The words were presented in lower case and Chicago font (size 80: red or blue for the study task and black for the test) on a light grey background and were displayed in the centre of a 17-inch flat monitor located at a distance of one metre from the participant. In the study task, the duration of each stimuli was 2500 ms, and a random inter-stimulus interval of between 800 and 1200 ms was used. In the test, each stimulus lasted up to 2000 ms (the word disappeared from the monitor screen when the participant pushed the response button), and the duration of the inter-stimulus interval was between 800 and 1200 ms. In the study task and the test, a black cross was presented in the centre of the monitor on a light grey background during the inter-stimulus interval, and the participants were instructed to keep their gaze fixed on the cross. The 159 words (of which 80 referred to living beings and 79 to nonliving items) were selected according to the following criteria: valence, arousal and familiarity of the word in the context of Spanish speech, degree of concretion, frequency of use and length. 2.3. EEG recording Participants were seated on a comfortable chair in an electrically shielded room, with attenuated levels of light and noise, and were instructed to move as little as possible during the recording. EEG activity was recorded with a Brain Vision Recorder device, via 60 active electrodes placed in an elastic cap (Easycap, GmbH), according to the International 10–10 system. All electrodes were referenced to an electrode attached to the tip of the nose and an electrode positioned at Fpz served as a ground. The horizontal electro-oculogram (HEOG) was recorded via two electrodes placed at the outer canthi of both eyes, whereas the vertical EOG (VEOG) was recorded via two electrodes placed supra and infraorbitally to the right eye. The EEG was continuously digitized at a rate of 500 Hz (bandpass filter 0.01–100 Hz) and electrode impedances were maintained below 10 kΩ. Once the signal was stored, the EEG data were exported to Matlab (R2017a version) and processed using EEGLAB (Delorme & Makeig, 2004) and ERPLAB (Lopez-Calderon & Luck, 2014). The signal was passed through a digital 0.1–30 Hz (12 dB/octave slope) bandpass filter, and ocular artefacts were corrected off-line by Independent Component Analysis (ICA). In order to study the brain activity related to test blocks of the immediate and delayed memory phases, the EEG was then segmented by extraction of epochs from −200 to 2000 ms post stimuli, only for correctly recognized old words and with correct colour recollection (Hit/Hit, H/H), and for correctly rejected new words (CR). All epochs were corrected to the mean voltage of 200 ms pre-stimulus recording period, and EEG segments exceeding ± 100 µV were automatically rejected. EEG corresponding to old words that were correctly recognized without posterior colour recollection (Hit/Miss, H/M) were not analyzed in the present study because the number of epochs for the averages in each phase (immediate and delayed) was insufficient (< 20 epochs). The mean number of averaged epochs (SD: standard deviation) for each phase was as follows: Immediate memory phase: 46.9 ( ± 9.4) for H/H and 65.6 ( ± 11.2) for CR; Delayed memory phase: 27.8 ( ± 5.2) for H/H and 34.6 ( ± 4.2) for CR. EEG related to each study block were not analyzed. 2.4. Procedure 2.4.1. Data analysis 2.4.1.1. Behavioural data. Reaction times (RT, between the onset of the word and pressing the key) and the percentage of responses were recorded in the following experimental conditions: H/H, H/M, CR, errors in the recognition of old words (as they were considered new words), false alarms (new words were considered old words) and misses (response omissions). In the present study we only evaluated the responses: H/H, H/M and CR. 2.4.1.2. ERPs data. The same ERP components were identified in the grand-average ERP waveforms of the immediate and delayed memory phases. All ERP waveforms included a positive followed by a negative peak at about 100 and 200 ms (P100 and N170 components) respectively, but only the ERP components identified from 200 ms onwards were analyzed. The associated parameters (amplitude and latency) were evaluated in different temporal ranges determined from the grand-average ERP waveforms. The following ERP components were evaluated: a negative component, with maximal amplitude at frontal scalp sites, in the 200–500 ms post-stimulus interval (coinciding with the temporal range described in the literature for the mid-frontal old/new effect or FN400 effect), and a positive component with maximal amplitude at parietal scalp sites, in the 400–900 ms post-stimulus interval (coinciding with the temporal range described in the literature for the parietal old/new effect). Mean amplitudes were measured for both components considering an interval around the peak, which was estimated at the electrode sites where the amplitude of each component was largest: at Fz for the frontal negative component, and at Pz for the parietal positive component. In the immediate memory phase, the mean amplitude of the frontal negative component was measured in the 250–350 ms interval (considering ± 50 ms around the peak), and the mean amplitude of the parietal positive component in the 500–650 ms interval (considering ± 75 ms M.Á. Rivas-Fernández, et al. Neurobiology of Learning and Memory 175 (2020) 107309 4 around the peak in the ERP waveforms to the old words). Similar measurements were made in the delayed memory phase, although with slightly different temporal intervals compared to the immediate memory phase due to slight differences in the latency of the peak amplitude: 270–370 ms ( ± 50 ms around the peak) for the frontal negative component and 525–675 ms ( ± 75 ms around the peak) for the parietal positive component. Mean amplitudes of both components were evaluated at the following electrode clusters (scalp regions of interest -ROIfor statistical analyses), previously employed in the literature (Voss & Federmeier, 2011): Mid frontal (F1, Fz, F2, FC1, FCz, FC2, C1, Cz, C2), mid posterior (CP1, CPz, CP2, P1, Pz, P2, POz), left posterior (TP7, CP5, CP3, P7, P5, P3, PO7, PO3) and right posterior (TP8, CP6, CP4, P8, P6, P4, PO8, PO4). In addition, the 50% fractional area latency was also estimated for both components. This involved computing the area under the ERP waveform over a given latency range and then finding the time point that divides the area into a 50% fraction (Luck, 2014). The latency range for calculation of the area of each component was established observing the grand-average ERP waveforms comprising the complete duration of the component (onset to end). Fractional area latency was evaluated for the frontal negative component and the parietal positive component at those electrodes where these components showed maximal amplitude, that is, Fz for the frontal negative component and Pz for the positive parietal component. In both memory phases of the task (immediate and delayed memory phases), the latency of the frontal negative component for both conditions (H/H and CR) was estimated in the 200–500 ms interval. In addition, the latency of the parietal positive component was estimated: (1) in the immediate memory phase, using different temporal intervals for the H/H (400–850 ms) and CR (400–900 ms) conditions, and (2) in the delayed memory phase, using the 350–900 ms interval for both conditions (H/H and CR). The LPN was identified at parieto-occipital locations, i.e. from 800 ms onwards in the immediate memory phase, and from 950 ms onwards in the delayed memory phase. In order to evaluate this component in both conditions (H/H and CR), mean amplitudes were also measured at the four ROIs aforementioned (mid frontal, mid posterior, left posterior and right posterior), in seven temporal windows (each of 150 ms) between 950 and 2000 ms post-stimuli, in the immediate and delayed memory phases. Voltage maps were also obtained for topographic analysis of the frontal negative component, the parietal positive component and the LPN, in each phase (immediate and delayed) and condition (H/H and CR). The voltage maps were calculated in the temporal ranges where the mean amplitudes of these components were identified. 2.4.1.3. Brain source localization analysis. Exact low-resolution electromagnetic tomography (eLORETA) software (publicly available, free academic software, at http://www.uzh.ch/keyinst/loreta.htm) was used to estimate cortical sources of EEG activity (recorded at 60 scalp electrodes) in the intervals where the frontal negative component, the parietal positive component, and the LPN were identified for the H/H and CR conditions during both phases of the task. This software is a three-dimensional, discrete, linear, weighted minimum norm inverse solution method that estimates the sources of EEG recorded on the scalp. A 3 shell-spherical head model is used as a reference (using the Talairach human brain atlas) and is divided in 6239 voxels at a spatial resolution of 5 mm. eLORETA represents the electrical activity at each voxel in neuroanatomical Montreal Neurological Institute (MNI) space, although Talairach coordinates, anatomical structures and Brodmann areas are also provided. For a more detailed description of the method and the exact zero-error localization property, see Pascual-Marqui et al. (2011). 2.5. Statistical analysis 2.5.1. Behavioural data Regarding the immediate memory phase, repeated measures analysis of variance (ANOVA) with two within-subject factors, Condition (three levels: H/H, H/M, CR) and Test Block (three levels: first, second and third), was conducted to compare the percentage of responses and the RTs among H/H, H/M, CR conditions and across the three test blocks. Regarding the delayed memory phase, a repeated measures ANOVA with one within-subject factor, Condition (three levels: H/H, H/M, CR), was used to compare the percentage of responses and the RTs among the three conditions. In addition, paired-sample t-tests were used to compare the RTs and the percentage of H/H responses, between the third test block of the immediate memory phase and the test of the delayed memory phase. 2.5.2. ERPs data Repeated measures ANOVAs with three within-subject factors, Phase (two levels: Immediate, Delayed), Condition (two levels: H/H, CR) and ROI (four levels: mid frontal, mid posterior, left posterior, right posterior), were conducted in order to compare, between phases and conditions and among ROIs, the mean amplitudes of the frontal negative component and the parietal positive component. In the case of the LPN, a repeated measures ANOVA with four within-subject factors Phase (two levels: Immediate, Delayed), Condition (two levels: H/H, CR), ROI (four levels: mid frontal, mid posterior, left posterior, right posterior) and Interval (seven levels: 950–1100 ms, 1100–1250 ms, 1250–1400 ms, 1400–1550 ms, 1550–1700 ms, 1700–1850 ms and 1850–2000 ms) was performed in order to compare the mean amplitude between phases and conditions, and among ROIs and temporal intervals. In addition, fractional area latency was evaluated via a repeated measures ANOVA with two within-subject factors, Phase (two levels: Immediate, Delayed) and Condition (two levels: H/H, CR) in order to compare, between phases and conditions, the latencies of the frontal negative component (at Fz) and the parietal positive component (at Pz). Greenhouse-Geisser corrections to the degrees of freedom were applied in all cases in which the condition of sphericity was not met. In these cases, the original degrees of freedom are presented together with the corrected p and ε values. When the ANOVAs showed significant effects of the factors and/or their interactions for the behavioural or ERP data, post-hoc testing of the mean values was carried out by paired multiple comparisons (with Bonferroni corrections). In addition, partial eta squared (η2p) was calculated for each significant comparison, with the aim of determining the size of the effects. Statistical significance was considered to be p ≤ 0.05. The statistical analyses were performed with IBM SPSS Statistics package v.21 for Windows. 2.5.3. Brain source localization analysis The eLORETA software package was used to perform (voxel-byvoxel) within-subject comparisons. This non-parametric method is based on estimation, via randomization, of the empirical probability distribution for the max-statistic under the null hypothesis, and it corrects for multiple testing (Nichols & Holmes, 2002). The following statistical comparisons were made: (1) H/H relative to CR conditions, in each memory phase (immediate and delayed), in order to compare the activation of neural sources between both conditions, and (2) immediate relative to delayed memory phases for each condition (H/H and CR), in order to compare activation of neural sources between both memory phases, for the correctly recognized old words (which were posteriorly accompanied by correct colour recollection, H/H) and for the correctly rejected new words (CR). These comparisons were performed (1) in both temporal ranges in which the frontal negative and the parietal positive components were identified: the 200–500 ms interval and 400–900 ms interval, respectively, and (2) in the 1200–1600 ms interval, around which the maximum amplitude differences between M.Á. Rivas-Fernández, et al. Neurobiology of Learning and Memory 175 (2020) 107309 5 conditions (H/H and CR) were observed for LPN, in based on the grandaverage ERP waveforms of the immediate memory phase (in which a significant LPN was obtained, see Results section). Statistical comparisons between conditions or phases, for the current density distribution, were estimated using paired sample t-test. Statistical significance was considered to be p ≤ 0.05. 3. Results 3.1. Performance The mean values of RTs and the percentage of responses in each condition (H/H, H/M, CR, errors to old words, false alarms and misses) are shown in Table 1. Mean values of RTs and the percentage of responses in H/H, H/M and CR are represented in Fig. 2. 3.1.1. Reaction times Regarding the immediate memory phase, the repeated-measures ANOVA (Test Block × Condition) showed a significant effect of the Test Block factor (F(2,30) = 26.4, p < 0.001, ε = 0.71; η 2 p = 0.63). Post hoc comparisons showed that the mean RT in the three conditions was significantly longer in the first than in the second (p = 0.002) and the third (p < 0.001) test blocks, and significantly longer in the second than in the third (p = 0.001) test block. No significant differences were observed for the RTs between the H/H, H/M and CR conditions in each phase, nor between the third test block of the immediate memory phase and the test of the delayed memory phase in the H/H condition. 3.1.2. Percentage of responses Regarding the immediate memory phase, the repeated measures ANOVA (Test Block × Condition) showed a significant effect of the Test Block factor (F(2,30) = 37.7, p < 0.001, ε = 0.72, η 2 p = 0.71), Condition factor (F(2,30) = 189.7, p < 0.001, ε = 0.61, η 2 p = 0.93) and of the Test Block × Condition interaction (F(4,60) = 11.3, p < 0.001, ε = 0.51, η 2 p = 0.43). Post hoc comparisons revealed that the percentage of CR and H/H responses was significantly lower in the first than in the second and third test blocks [CR first vs second: p = 0.03, first vs third: p < 0.001] [H/H first vs second: p = 0.003, first vs third: p = 0.001], and the percentage of CR responses was also significantly lower in the second than in the third test block (p = 0.008). In addition, the percentage of H/M responses was significantly higher in the first than in the third test block (p = 0.03). In the three blocks, the percentage of CR responses was significantly higher than the percentage of H/H and H/M responses [CR versus H/H in first block: p < 0.001, second block: p < 0.001, third block: p < 0.001] [CR versus H/M in the first block: p < 0.001, second block: p < 0.001, third block: p < 0.001], and the percentage of H/H responses was significantly higher than the percentage of H/M Table 1 Mean values and standard deviations (between parentheses) for the percentage of responses (%) and the reaction times (RTs, ms), in each condition: Hit/Hit (H/H), Hit/Miss (H/M) and Correct Rejection (CR), errors to old words, false alarms and misses. Immediate Phase Delayed Phase Test-Block 1 Test-Block 2 Test-Block 3 Mean Test Blocks Delayed test % RT % RT % RT % RT % RT H/H 54.8 (11.5) 872 (129) 68.7 (9.5) 792 (140) 75.0 (13.7) 722 (136) 66.2 (8.4) 796 (124) 75.8 (11.0) 698 (110) H/M 31.7 (14.6) 896 (215) 23.3 (11.3) 795 (156) 19.4 (12.3) 716 (167) 24.8 (10) 802 (145) 18.7 (9.8) 718 (142) CR 86.5 (9.6) 921 (141) 92.8 (8.3) 789 (121) 98.2 (3.8) 705 (98) 92.5 (6.2) 805 (102) 97.0 (3.1) 686 (101) ERRORS TO OLD WORDS 6.7 (8.4) 857 (219) 3.9 (5.3) 1084 (412) 2.8 (4.6) 769 (200) 4.5 (5.6) 873 (256.4) 2.8 (4.7) 783 (164) FALSE ALARMS 7.2 (7.1) 1067 (287) 2.9 (3.6) 766 (202) 0.2 (0.9) – a 3.4 (3.2) – b 1.9 (2.5) 775 (210) MISSES TO OLD WORDS 6.7 (9.1) 4.1 (4.3) 2.8 (4.2) 4.5 (4.6) 2.7 (4.5) MISSES TO NEW WORDS 6.3 (10) 4.3 (8.5) 1.6 (3.8) 4.1 (7.0) 1.1 (2.7) a Mean and standard deviation of the mean reaction time to false alarms in the third test block were not calculated because there was only one participant with one false alarm in that block of the task. b Mean and standard deviation of the mean reaction times to false alarms was not calculated because most of participants did not have false alarms in one or more test blocks. Fig. 2. Mean values of the percentage of responses (left) and reaction times (RTs, right) in each condition: correct recognition of old words with retrieval of their colour (Hit/Hit), correct rejection of new words (CR) and correct recognition of old words without retrieval of their colour (Hit/Miss). IMP-1: Immediate memory phase-Test 1; IMP-2: Immediate memory phase-Test 2; IMP-3: Immediate memory phase-Test 3; DMP: Delayed memory phase test. **p < 0.01, ***p < 0.001. M.Á. Rivas-Fernández, et al. Neurobiology of Learning and Memory 175 (2020) 107309 6 responses (first block: p = 0.006, second block: p < 0.001, third block: p < 0.001). Regarding the delayed memory phase, the repeated measures ANOVA (Condition) revealed a significant effect of the factor (F(2,30) = 248.7, p < 0.001, η 2 p = 0.9), as the percentage of CR responses was significantly higher than the percentage of H/H (p < 0.001) and H/M (p < 0.001) responses, and the percentage of H/H was significantly higher than the percentage of H/M responses (p < 0.001). The percentage of H/H responses was not significantly different between the third test block of the immediate memory phase and the test of the delayed memory phase. 3.2. Event-related potentials (ERPs) Grand-average ERP waveforms for the H/H and CR conditions at mid frontal, mid posterior, left posterior and right posterior ROIs in the immediate and delayed memory phases, are shown in Fig. 3, and the voltage maps are shown in Fig. 4. Mean amplitude and fractional latency results for the frontal negative component and the positive parietal component are summarized in Table 2. Mean amplitude results for the LPN are summarized in Table 3. For the amplitude of the frontal negative component, the repeated measures ANOVA (Phase × Condition × ROI) showed a significant Fig. 3. Grand-average ERP waveforms at the mid frontal, mid posterior, left posterior and right posterior electrode clusters in the H/H and CR conditions, during the immediate memory phase and delayed memory phase. The red and blue dotted lines represent the mean RT of the H/H and CR conditions, respectively. The black dashed line represents the mean time of the presentation of the colour question. The grey shaded area represents intervals with significant results. M.Á. Rivas-Fernández, et al. Neurobiology of Learning and Memory 175 (2020) 107309 7 effect of the ROI factor and the Phase × ROI interaction. Post hoc comparisons revealed that, during the immediate memory phase, the amplitude was significantly more negative at the mid frontal than at the mid posterior (p = 0.007) and the right posterior (p = 0.047) ROIs. In addition, during the delayed memory phase, it was significantly more negative at the mid frontal than at the mid posterior (p = 0.003), the left posterior (p = 0.044) and the right posterior (p = 0.008) ROIs. The repeated measures ANOVA (Phase × Condition) did not show any significant effect of the factors or their interaction for the latency of the frontal negative (FN400) component. For the amplitude of the parietal positive component, the repeated measures ANOVA (Phase × Condition × ROI) showed a significant effect of the Phase, Condition and ROI factors as well as the Phase × Condition and the Phase × ROI interactions. Post hoc comparisons revealed that the mean amplitude was significantly larger: (1) in the H/H than in the CR condition in both memory phases (immediate: p = 0.005, delayed: p < 0.001), indicating a significant parietal old/new effect; (2) at the mid posterior than at the mid frontal ROI (p = 0.005) during the immediate memory phase, (3), at the mid posterior than at the mid frontal (p < 0.001) and the right posterior (p = 0.01) ROI, and at the left posterior than at the mid frontal ROI (p = 0.02), during the delayed memory phase; (4) in the delayed than in the immediate memory phase (p = 0.02), in the H/H condition; and (5) in the delayed than in the immediate memory phase in the mid posterior (p = 0.03) and left posterior (p = 0.03) ROIs. For the parietal positive component latency, the repeated measures ANOVA (Phase × Condition) showed a significant effect of the Fig. 4. Voltage maps for the mean amplitude of the frontal negativity component (top), the parietal positive component (middle) and the late posterior negativity (LPN, bottom), in the H/H and CR conditions, during the immediate and delayed memory phases. Table 2 F values in repeated measures ANOVA (Phase × Condition × ROI) for the mean amplitudes and 50% fractional area latencies of the frontal negative component and the parietal positive component. Frontal negative component Parietal positive component Mean Amplitude ANOVA (P × C × R) P 1.4, η 2 p = 0.09 4.6*, η 2 p = 0.24 C 4.4, η 2 p = 0.23 24.2***, η 2 p = 0.62 R 7.3***, η 2 p = 0.33 11.2***, η 2 p = 0.43 P X C 1.4, η 2 p = 0.08 6.5*, η 2 p = 0.30 P X R 5.1*, ε = 0.56; η 2 p = 0.25 10.8***, η 2 p = 0.42 C X R 0.8, ε = 0.52; η 2 p = 0.05 2.8, ε = 0.71; η 2 p = 0.16 P X C X R 0.1, ε = 0.51; η 2 p = 0.006 2.4, ε = 0.56; η 2 p = 0.14 Fractional Area Latency ANOVA (P × C) P 2.2, η 2 p = 0.13 0.003, η 2 p < 0.001 C 0.7, η 2 p = 0.04 11.2**, η 2 p = 0.43 P × C 3.6, η 2 p = 0.2 0.95, η 2 p = 0.06 P: Phase factor; C: Condition factor; R: ROI factor; ε = epsilon, η 2 p = partial eta squared value. Degrees of freedom for: P: 1,15; C: 1,15; R: 3,45 and the interactions P × C: 1,15; P × R: 3,45; C × R: 3,45; P × C × R: 3,45. *p < 0.05; **p < 0.01; ***p < 0.001 Table 3 F values in repeated measures ANOVA (Phase × Condition × ROI × Interval) for the mean amplitudes of the LPN in seven temporal windows of 150 ms. Late Posterior Negativity (LPN) Mean Amplitude ANOVA (P × C × R × I) P 9.4**, η 2 p = 0.39 C 12.1**, η 2 p = 0.45 R 2.6, η 2 p = 0.15 I 2.4, η 2 p = 0.14 P × C 10.6**, η 2 p = 0.41 P × R 1.7, η 2 p = 0.10 P × I 0.8, η 2 p = 0.47 C × R 3.7*, ε = 0.55, η 2 p = 0.20 C × I 8.8***, ε = 0.42, η 2 p = 0.37 R × I 2.1**, η 2 p = 0.12 P × C × R 2.9, η 2 p = 0.16 P × C × I 0.5, η 2 p = 0.03 P × R × I 1.5, η 2 p = 0.09 C × R × I 6.5***, η 2 p = 0.30 P × C × R × I 1.1, η 2 p = 0.07 P: Phase factor; C: Condition factor; R: ROI factor; I: Interval; ε = epsilon, η 2 p = partial eta squared value. Degrees of freedom for: P: 1,15; C: 1,15; R: 3,45; I: 6,90 and the interactions P × C: 1,15; P × R: 3,45; C × R: 3,45; P × C × R: 3,45; P × I: 6,90, C × I: 6,90, P × C × I: 6,90, R × I: 18,270, P × R × I: 18,270, C × R × I: 18,270, P × C × R × I: 18,270. *p < 0.05; **p < 0.01; ***p < 0.001. M.Á. Rivas-Fernández, et al. Neurobiology of Learning and Memory 175 (2020) 107309 8 Condition factor. Post hoc comparisons revealed that the latency was significantly shorter in the H/H than in the CR condition (p = 0.004). For the LPN mean amplitude, the repeated measures ANOVA (Phase × Condition × ROI × Interval) showed a significant effect of the Phase and Condition factors, as well as the following interactions: Phase × Condition, Condition × ROI, Condition × Interval, ROI × Interval, and Condition × ROI × Interval. Post hoc comparisons revealed that the mean amplitude of LPN was significantly more negative in the H/H condition than in the CR condition (p = 0.001) in the immediate memory phase; and no significant differences were obtained between both conditions (p = 0.07) in the delayed memory phase, although the differences were in the same direction (H/H more negative than CR). In addition, the mean amplitude of the H/H condition was significantly more negative in the immediate than in delayed memory phase (p = 0.002). On the other hand, post hoc comparisons also revealed that the mean amplitude in the H/H condition was significantly more negative than the CR condition, at the following ROIs and intervals: (1) Mid frontal locations, in the 950–1100 ms (p = 0.004), 1100–1250 ms (p = 0.001), 1250–1400 ms (p = 0.001), 1400–1550 ms (p < 0.001), 1550–1700 ms (p = 0.002), 1700–1850 (p = 0.017) intervals; (2) Mid posterior locations, in the 950–1100 ms (p = 0.001), 1100–1250 ms (p < 0.001), 1250–1400 ms (p < 0.001), 1400–1550 ms (p = 0.001), 1550–1700 ms (p = 0.002), 1700–1850 (p = 0.049) intervals; (3) Left posterior locations, in the 950–1100 ms (p = 0.017), 1100–1250 ms (p = 0.002), 1250–1400 ms (p = 0.002), 1400–1550 ms (p = 0.003), 1550–1700 ms (p = 0.007) intervals and (4) Right posterior locations, in the 950–1100 ms (p = 0.001), 1100–1250 ms (p = 0.001), 1250–1400 ms (p = 0.001), 1400–1550 ms (p = 0.005) and 1550–1700 ms (p = 0.018) intervals. In addition, post hoc comparisons revealed that the mean amplitude of LPN in the H/H condition was significantly more negative: (1) at the left posterior than at the right posterior locations in the 1250–1400 ms interval (p = 0.031) and the 1850–2000 ms interval (p = 0.041), and (2) at the mid posterior and the left posterior than the right posterior locations in the 1550–1700 ms interval (p = 0.002 and p = 0.016, respectively) and in the 1700–1850 ms interval (p = 0.037 and p = 0.012, respectively). The mean amplitude of LPN in the CR condition was significantly more negative: (1) at the left posterior than at the mid posterior locations during the 950–1100 ms (p = 0.018) and the 1100–1250 ms (p = 0.005) intervals, (2) at the mid frontal (p = 0.049) and at the left posterior (p = 0.003) than at the mid posterior locations, as well as at the left posterior than at the right posterior locations (p = 0.05), during the 1250–1400 ms, and (3) at the left posterior than at the mid posterior locations in the 1400–1550 ms (p = 0.006) and the 1550–1700 ms intervals (p = 0.039). Finally, post hoc comparisons also revealed that the mean amplitude of LPN in the H/H condition was significantly more negative: (1) in the 1550–1700 ms (p = 0.018) and 1700–1850 ms (p = 0.016) intervals than in the 1850–2000 ms interval, at the mid posterior locations, (2) in the 1100–1250 ms (p = 0.02), 1550–1700 ms (p = 0.014) and 1700–1850 ms (p = 0.013) intervals than in the 1850–2000 ms interval, at the left posterior locations, and (3) in the 950–1100 ms (p = 0.002), 1100–1250 ms (p = 0.008), 1400–1550 ms (p = 0.037), 1550–1700 ms (p = 0.004) and the 1700–1850 ms (p = 0.007) intervals than in the 1850–2000 ms interval, as well as in the 1550–1700 ms than in the 1700–1850 ms interval (p = 0.034), at the right posterior locations. 3.3. Brain source localization analysis Table 4 shows the MNI coordinates and Brodmann areas of those brain regions in which significant difference in activation was observed in the following: (1) the H/H > CR contrast, during the immediate and delayed memory phases; and (2) the delayed memory phase > immediate memory phase contrast, in the H/H and CR conditions. During the immediate memory phase, the paired sample t-test did not reveal any significant differences in brain activation between the H/H and CR conditions in the temporal interval of the frontal negative component (200–500 ms); however, they showed significantly greater neural activity in the H/H than the CR condition in the parietal positive component interval (400–900 ms) [t (15) = 4.1, p < 0.05], specifically in the inferior parietal lobule (BA: 40), parahippocampal gyrus (BA: 28, 34), uncus (BA: 28, 34), precuneus (BA: 7) and lingual gyrus (BA: 18) of the left hemisphere, and in the right posterior cingulate (BA: 30). In the temporal interval of LPN (1200–1600 ms), the paired sample t-test showed significantly greater neural activity in the H/H than the CR condition [t (15) = 3.8, p < 0.05] in the parahippocampal gyrus (BA: 28), anterior cingulate gyrus (BA: 32), insula (BA: 13), superior, medial and middle frontal gyri (BA: 10, 32, 46), uncus (BA: 28) and transverse temporal gyrus (BA: 41) of the left hemisphere, as well as in the rectal gyrus (BA: 11), inferior frontal gyrus (BA: 46), subcallosal gyrus (BA: 25) and posterior cingulate gyrus (BA: 30) of the right hemisphere (see Table 4 and Fig. 5). Regarding the delayed memory phase, the paired sample t-test did not reveal any significant differences in brain activation between the H/H and CR conditions in the frontal negative component interval (200–500 ms). In the parietal positive component interval (400–900 ms), the paired sample t-test showed significantly greater neural activity for the H/H than the CR condition [t (15) = 4.0, p < 0.05] in the insula (BA: 13), posterior cingulate gyrus (BA: 31), uncus (BA: 28), precuneus (BA: 7) and parahippocampal gyrus (BA: 34) of the left hemisphere, the bilateral superior temporal gyrus (BA: 22, 42) and also the posterior cingulate gyrus (BA: 23) of the right hemisphere. In the LPN temporal interval (1200–1600 ms), the paired sample t-test showed significantly greater neural activity for the H/H than the CR condition (t (15) = 3.9, p < 0.05) in the superior temporal gyrus (BA: 41) and insula (BA: 13) of the left hemisphere, and also in the right parahippocampal gyrus (BA: 28) (see Table 4 and Fig. 5). Finally, within the H/H condition, the paired sample t-tests revealed significantly greater activation [t (15) = − 3.9, p < 0.05)] of the left insula (BA: 13) and the right medial frontal gyrus (BA: 10) in the delayed compared to the immediate memory phase during the LPN temporal interval (1200–1600 ms) (see Table 4 and Fig. 6). In addition, for the CR condition, significantly greater activation [t (15) = − 4.4, p < 0.01)] of the right parahippocampal gyrus (BA: 19) was observed in the delayed memory phase than in the immediate memory phase, during the temporal interval of LPN (1200–1600 ms) (see Table 4 and Fig. 6); however, no differences between both memory phases were observed in the 400–900 ms interval. 4. Discussion In the present ERP study using the exact low electromagnetic resolution tomography (eLORETA), we characterized and compared the performance and the spatiotemporal pattern of the brain electrical activity related to the successful immediate episodic retrieval of information which is being learned relative to delayed episodic retrieval (twenty minutes later) of the same information. For this purpose, young participants carried out an old/new word recognition task with source memory, which included an immediate memory phase (with three study-test blocks) followed (after 20 min) by a delayed memory phase with one test block. Behavioural data showed gradual learning and consolidation of the information (old words) during the immediate memory phase. The ERP data showed that, relative to correctly identified new words (CR responses), correctly judged old words (H/H responses) elicited the following in both memory phases: (1) a more positive-going potential, demonstrating a parietal old/new effect (around 500–675 ms post-stimulus interval) related to recollection, and (2) a more negative-going potential, demonstrating an LPN effect (in the 950–1850 ms post-stimulus interval at the mid frontal and M.Á. Rivas-Fernández, et al. Neurobiology of Learning and Memory 175 (2020) 107309 9 Vilberg, K. L., & Rugg, M. D. (2007). Dissociation of the neural correlates of recognition memory according to familiarity, recollection, and amount of recollected information. Neuropsychologia, 45(10), 2216–2225. https://doi.org/10.1016/j. neuropsychologia.2007.02.027. Vilberg, K. L., & Rugg, M. D. (2008). Memory retrieval and the parietal cortex: A review of evidence from a dual-process perspective. Neuropsychologia (Vol. 46(7), 1787–1799. https://doi.org/10.1016/j.neuropsychologia.2008.01.004. Voss, J. L., & Federmeier, K. D. (2011). FN400 potentials are functionally identical to N400 potentials and reflect semantic processing during recognition testing. Psychophysiology, 48(4), 532–546. https://doi.org/10.1111/j.1469-8986.2010. 01085.x. Wagner, A. D., Shannon, B. J., Kahn, I., & Buckner, R. L. (2005). Parietal lobe contributions to episodic memory retrieval. Trends in Cognitive Sciences (Vol. 9(9), 445–453. https://doi.org/10.1016/j.tics.2005.07.001. Wang, T. H., de Chastelaine, M., Minton, B., & Rugg, M. D. (2012). Effects of age on the neural correlates of familiarity as indexed by ERPs. Journal of Cognitive Neuroscience. https://doi.org/10.1162/jocn_a_00129. Wilding, E. L., Fraser, C. S., & Herron, J. E. (2005). Indexing strategic retrieval of colour information with event-related potentials. Cognitive Brain Research, 25(1), 19–32. https://doi.org/10.1016/j.cogbrainres.2005.04.012. Wilding, Edward L., & Rugg, M. D. (1996). An event-related potential study of recognition memory with and without retrieval of source. Brain, 119(3), 889–905. https://doi. org/10.1093/brain/119.3.889. Winocur, G., & Moscovitch, M. (2011). Memory transformation and systems consolidation. Journal of the International Neuropsychological Society: JINS, 17(5), 766. https:// doi.org/10.1017/S1355617711000683. Wolk, D. A., Schacter, D. L., Lygizos, M., Sen, N. M., Chong, H., Holcomb, P. J., ... Budson, A. E. (2007). ERP correlates of Remember/Know decisions: Association with the late posterior negativity. Biological Psychology, 75(2), 131–135. https://doi.org/10.1016/ j.biopsycho.2007.01.005. Wolk, D. A., Schacter, D. L., Lygizos, M., Sen, N. M., Holcomb, P. J., Daffner, K. R., & Budson, A. E. (2006). ERP correlates of recognition memory: Effects of retention interval and false alarms. Brain Research. https://doi.org/10.1016/j.brainres.2006. 04.050. Wolk, D. A., Sen, N. M., Chong, H., Riis, J. L., McGinnis, S. M., Holcomb, P. J., & Daffner, K. R. (2009). ERP correlates of item recognition memory: Effects of age and performance. Brain Research. https://doi.org/10.1016/j.brainres.2008.11.014. Woodruff, C. C., Hayama, H. R., & Rugg, M. D. (2006). Electrophysiological dissociation of the neural correlates of recollection and familiarity. Brain Research, 1100(1), 125–135. https://doi.org/10.1016/j.brainres.2006.05.019. Yonelinas, A. P. (2002). The Nature of Recollection and Familiarity: A Review of 30 Years of Research* 1. Journal of Memory and Language, 46(3), 441–517. https://doi.org/10. 1006/jmla.2002.2864. Yonelinas, A. P., Otten, L. J., Shaw, K. N., & Rugg, M. D. (2005). Separating the brain regions involved in recollection and familiarity in recognition memory. The Journal of Neuroscience, 25(11), 3002. http://neuro.cjb.net/content/25/11/3002.short. Yu, S. S., & Rugg, M. D. (2010). Dissociation of the electrophysiological correlates of familiarity strength and item repetition. Brain Research. https://doi.org/10.1016/j. brainres.2009.12.071. M.Á. Rivas-Fernández, et al. Neurobiology of Learning and Memory 175 (2020) 107309 16