An ERP study into the effects of aging on the
adoption of retrieval orientation with preparatory
cue processes.
Maxime S.X. KrausMaster Brain and Cognitive Sciences University of Amsterdam
27/08/2017 maxime.kraus@gmail.com
10216626
Supervisor: Dr. Alexa Morcom Co-assessor: Dr. Jan Engelmann
Abstract
Decline in recollection is often related to healthy cognitive aging, which seems also associated with impairment in self-initiation deficit, cognitive control and working memory. Older adults seem to have more difficulty internalising retrieval goals and rely more on external cues for retrieval. The encoding of these pre-retrieval processes may shed a light on the decline in recollection. In this study the preparatory processes before retrieval are investigated, in particular on the adoption and maintenance of retrieval orientations (e.g. processes helping towards the retrieval goal). It is hypothesised that older adults have a difficulty with using the preparatory cues for adoption of retrieval orientation. In a paradigm similar to the one in Herron, Wilding, and Rugg (2016),
participants were expected to switch predictably between which retrieval task to retrieve depending on the preparatory cue presented. The behavioural results suggested that older adults have more difficulty with the switch trials than the stay trials, but the ERP data did not. This difference in effects could be attributed to the noisy EEG or the difficulty in the task. However, the results do suggest an age effect for the preparatory cue processes on switch trials.
Keywords: Event-related potentials, Episodic memory, Cue Processing, Retrieval Orientation, Preparatory processing
Introduction
By 2050, 20% of the population will be above 60 years of age (WHO, 2015). Understanding the changes in cognition in a healthy aging brain is crucial in order for society to help and accommodate increasing number of older adults. One of the changes in normal cognitive aging is the decline in episodic memory (e.g. the memory of experienced events), as well as the decline in the ability to recall details, associations, and contextual information (Old & Naveh-Benjamin, 2008; Spencer & Raz, 1995). The tasks most affected by age seem to be tasks involving deliberate control over retrieval (Light, 1991). In Craik and McDowd (1987), they discovered that aging effects were greatest on free recall (no cues given), reduced in cued, absent when participants had to recognise the item. This seems to suggest that older adults have a self-initiation deficit (Craik, 1983). This pattern could be explained by the theory that older adults seem to have difficulty recollecting specific details of experienced events, while their sense of familiarity stays mainly intact (Koen & Yonelinas, 2014; Yonelinas, 2002). Overall, it seems that aging has an effect on the retrieval of memory, but it is less clear why recollection fails (Morcom, 2016). Deficits in cognitive control and working memory may be part of the explanation for the failure of recollection (Braver & West, 2008; Morcom, 2016). Processes prior to retrieval may also be critical for the failure of recollection. However, it also remains unclear how encoding before retrieval is affected by aging especially in preparatory processes. These pre-retrieval processes - processes activated by a pre-retrieval cue during a pre-retrieval attempt - have been relatively neglected in aging studies (Rugg &Wilding, 2000). These processes before retrieval are important to investigate aging’s effect on retrieval (Rugg & Morcom, 2005; Morcom, 2016). Recent research suggest that processes operating before retrieval are an important component of memory decline in age. Pre-retrieval processes such as preparatory processes and preretrieval control are important to understand first before improvement of retrieval can happen. These pre-retrieval processes are usually involved in targeting relevant mnemonic information prior to retrieval (Morcom, 2016). Cues can provide contextual information that was incorporated in the memory trace, which can trigger retrieval when memory trace is matched to the cue (Tulving, 1983; Tulving, 2002). These cues do not have to be external, but can also be internal by “self-cueing” or elaborating on external cues (Burgess, 1996). These cues could influence the recollection via the cue bias
mechanisms, which could influence the way retrieval cues are processed (Anderson, Bjork & Bjork, 1994). In this way, people can adopt a specific retrieval orientation that reflects retrieval goals. Neuroimaging methods such as electroencephalography (EEG) or functional magnetic resonance imaging (fMRI) can offer an insight into the preretrieval processes, because these processes can be measured without confounding pre-retrieval processing and recollection itself.
Retrieval orientations are considered pre-retrieval attentional states, which vary according to the type of episodic information retrieved (Rugg & Wilding, 2000). Retrieval orientations are usually examined by comparing physically identical retrieval cues under different retrieval goals (Johnson, Kounios & Nolde, 1997; Mecklinger, 2010; Morcom, 2016). These stimuli in retrieval tasks are considered to be episodic retrieval cues because they determine the processes needed to retrieve the retrieval goal (Dzulkifli, & Wilding, 2005). Usually retrieval orientation is measured by contrasting the
electroencephalographic event-related potentials (ERPs) evoked by new (unstudied) test items across tasks having different retrieval demands. This method is most effective, because it eliminates the possibility of indices of successful episodic retrieval (Herron, Evans & Wilding, 2016). The ERPs evoked by new items do vary according to retrieval demands, and these differences are not simply a
reflection of reaction time or difficulty changes across tasks. The ERP left parietal (LP) old/new effect is an ERP effect that occurs around 500-800ms and is positive going elicited by correctly identified studied items compared to unstudied items. This is considered to be recollection-driven recognition memory (Friedman & Johnson, 2000; for review see Rugg & Curran, 2007).
Most of the research is conducted on retrieval orientation (Morcom, 2016). Morcom and Rugg (2004) investigated the effect of aging on retrieval orientation, in specific, the differential processing of retrieval cues. Younger and older participants were presented with words or pictures. The ERPs elicited by unstudied (new) words were more positive going for the young than the old, the effects were more attenuated and delayed in older subjects. These findings were reflected in similar findings by Jacoby, Shimizu, Velanova, and Rhodes (2005a), which suggest that aging has an effect on the differential processing of retrieval cues. Duverne, Motamedinia and Rugg (2009) tried to investigate whether the retrieval cue processing in older adults was sensitive to retrieval task, specifically for age effects on cue processing strategy rather than in encoding. In this study, participants studied either pictures or words (same as in Morcom & Rugg, 2004), but one condition of the task was that participants had to remember the screen location (source information). With source information condition, older adults are capable of adopting retrieval orientation. They demonstrated that age effects are sensitive to the demands of the retrieval task, because when explicitly instructed older adults were able to recollect source information. This seems to cooperate with the belief that older adults have a self-initiation deficit, because unless specifically instructed no retrieval orientation was adopted.
However, research into the adoption of retrieval orientation could be important to investigate the self-initiation deficit in older adults, because it could offer a better method of investigating the deficit as well as perhaps provide a better explanation for the deficit. In cued task-switching paradigms as described by Herron and Wilding (2006), young adults seem to enter preparatory attention states, which can help orient information needed to be retrieved (Herron, et al, 2016; Herron & Wilding, 2006). Depending on the retrieval task, the preparatory neural activity varies, which is thought to indicate initiation of retrieval orientations (Herron & Wilding, 2006). Herron,et al.,(2016) discovered distinct task-dependent retrieval cue processing, which suggests that participants can flexibly adjust task-dependent retrieval cue processes. Preparatory cue processes can help orient the retrieval orientation in a similar manner as cue bias. By providing the cue, participants could already start searching for the memory trace as proposed by Tulving (1983). These processes are seen especially on the so called switch trials (trials in which the cues varies different retrieval goals after which they are immediately switched), when the retrieval orientation has to switch, because of the different retrieval goals of the task. Preparatory processes seem to be important to adopt retrieval orientation. It is important to investigate the adoption of retrieval orientation, because adoption of retrieval information is important for retrieval of relevant information from memory and stops irrelevant information from entering (Jacoby, Shimizu, Daniels & Rhodes, 2005b). However, preparatory neural activity in aging has not been investigated as much. A recent fMRI study (Dew, Buchler, Dobbins, & Cabeza., 2012) did investigate the preparatory activity by using a cue and probe paradigm. It was discovered that memory recovery in older adults was in response to the probe and not the cue, which indicates a self-initiation deficit. The posterior parietal cortex (PPC) had a greater probe-related activity in older adults. There was reduced preparatory activity in the hippocampus prior to episodic retrieval in older adults, which also supports the age impaired self-initiation deficit.
As discussed before, it seems that the self-initiation deficit could be attributed to the reduction in attentional sources, such as attentional control. Recent studies of attentional control have discovered particular functions affected by aging. In Braver’s dual mechanisms of control (DMC), two types of attentional control are discussed, the proactive control and the reactive control. Proactive control is necessary for self-initiated behaviour and is closely related to the ability to maintain goal information online. Reactive control is active after the interference has been present to resolve and is stimulus-driven and transient (Braver, 2012; Braver & Barch, 2002).
In a task, such as the AX Continuous Performance Task (AX-CPT), used to investigate attentional control, older adults seem to be impaired in proactive control and appear to rely on reactive control (Bugg, 2014). This shift from proactive control to reactive control could reflect in the failures of
pre-retrieval control in older adults. In the AX-CPT, participants are presented with a cue and probe pairs, depending on the cue and probe combination, a different response is expected. AX trials are the dominant trials, which bias the participants towards a higher response tendency to answer when the cue “A” is followed by the probe “X”. When participants use proactive control, the performance will be better on the BX trials (e.g any cue other than A) because the tendency to answer is diminished, while a participant using reactive control will respond more depending on the probe, thus, will perform worse on the BX trials. However, the opposite is true for AY trials (e.g. probe is any other letter than X), where proactive control is biased to answer and needs reactive control to override the response.
As mentioned before, proactive control is also closely related to the ability to maintain goal
information online, specifically in the working memory when interference is present. This is reflected in working memory capacity (WMC, Engle & Kane, 2003). It seems that the active goal maintenance has an influence on the episodic retrieval deficits. Participants with a lower WMC recalled fewer items than higher WMC participants. It is thought to be an effect that low WMC participants search through a larger set of items (Engle & Kane, 2003). Elward and Wilding (2010) discovered larger LP old/new effect for Targets in young participants with a higher WMC. However, Keating, Affleck-Brodie, Wiegand and Morcom (2017) discovered that young adults with a higher WMC show, as examined by the operation span test (O-span), greater recollection selectivity than young adults with lower WMC. In older adults, there seems to be an uncoupling of WMC and recollection selectivity.
This present study examined the differences in preparatory processes, especially focused on the adaptation and maintenance of retrieval orientations once adopted, due to aging in a modified version of the one used in Herron, et al.,(2016). In our study, there were different retrieval tasks and no unstudied words were used in the test. Participants performed a semantic encoding task (e.g. encoding is processing at the time of an event and is accessible later on in memory task), in which participants had to remember the source of the decisions made on the word (e.g. Artist or Function). In this encoding task, participants rated whether it was easy or difficult to draw something and how many functions an item could have, a few or many. In the test, participants were ‘prepared’ to retrieve the decision via preparatory cues which preceded each test item. Then they had to remember whether they made an Artist or a Function decision on the test item. These cues would switch predictably every two trials. We hypothesised that normal cognitive aging impairs the use of preparatory cues to adopt retrieval orientation. The preparatory cue ERPs on switch trials will be less attenuated with a delayed onset and earlier offset in older adults than in younger adults.
Older adults will not engage in pre-retrieval control in the simple recognition task, because it is not explicitly required (Duverne et al., 2009). It is also assumed that older adults will have an attenuated ERP effect on the stay trials, because they have difficulty maintaining the retrieval orientation.
Material & Methods Participants
Participants were 30 healthy younger (6 males; ages 20-29, M=22.77, SD=2.687) and 20 older (10 males; ages 67-76, M =69.60, SD= 2.521) right-handed adults with normal or corrected-to-normal vision, who all were native English speakers (see Table 1). Young participants were recruited through the University of Edinburgh student community (MyCareerHub) and older participants from the Psychology Volunteer Panel of adults (http://www.ed.ac.uk/ppls/psychology/research/volunteering) and the University of the Third Age (https://u3asites.org.uk/edinburgh/welcome). Data from five participants were excluded due to insufficient trials following EEG artefact removal or a near-chance performance defined by discrimination below 0.1 (e.g. discrimination between Target and Non-target trials; Target trials mean that the encoding task designation matched the cue, e.g. Artist task matched
Artist cue; Non-target trials were not matched on cue and encoding task designation). Participants took part in two experimental sessions for which an honorarium of £15 was paid and an informed consent was signed before participation. The study was approved by the Psychology Research Ethics Committee (ref.: 89-1617/7).
Methods
Participants were asked to provide their educational level. These levels were converged into the following levels: Level 1 = GSCE/Standard, Level 2 = A-levels/Higher, Level 3 = Diploma not a degree/ DipCE, level 4 = Undergraduate, Level 5 = Post-graduate Masters, Level 6 = PhD or equivalent (e.g. medical doctorate).
Standardised Cognitive Tests
To characterise the two age groups, we also administered several baseline tests next to two tasks to measure the WMC (O-span) and the proactive control (AX-CPT). The Test of Premorbid Functioning (TOPF) was used in combination with the Wechsler Adult Intelligence Scale III (WAIS-III) Vocabulary to assess crystallised verbal intelligence (Wechsler, 1997; Wechsler, 2011), which is expected to remain or even improve with age (Li et al., 2004). The TOPF was converted in a verbal IQ score. Processing speed was measured using the Digit-Symbol Coding subtest of the Wechsler Adult Intelligence Scale IV (WAIS IV; Wechsler, 2008), which is believed to decrease with age (Li et al., 2004; Salthouse, 2000). Associative memory was measured with the Verbal Paired Associates WAIS-IV Memory Scale (VPA; Wechsler, 2008), which is assumed to be worse in older adults, especially on the free recall,
immediate and delayed recall (Cullum, Butters, Tröster, & Salmon, 1990; Craik, 1984). In this version of the VPA, older adults received less items to remember (10 vs 14 for young). Therefore, the VPA scores are scaled in accordance to the Wechsler Memory Test Manual. These tests together with the AX-CPT and the O-span were administered in separate sessions that took about one hour and fifteen minutes. For analysis, the raw scores for the Coding, Verbal Paired Associates Recognition were used.
AX-CPT
To measure proactive control as discussed in the DMC, a modified version of the AX Continuous Performance Task (AX-CPT) was used in which participants respond to certain probe letters depending on the cue letters received (see figure 1; Braver, Paxton, Locke, & Barch2009). Each trial started with a cue letter presented for 500ms, after which a fixation cross for 3750ms, followed by the probe for 500ms, after which three fixation crosses were presented for 1000ms. Participants were given 1000ms after the onset of the cue and the probe to respond. On the AX pairs, Target responses were expected, while on the other pairs Non-target responses were made. The proportions of the trial types were based on those used by Richmond, Redick, and Braver., (2015): 40% of the trials consisted of an A followed by an X (AX trials), 10% of the trials were A followed by a letter other than X
(pseudo-randomly selected; AY trials), the other 10% were any letter followed by an X (BX trials), and 40% of the trials consisted of any letter (not an A) and any other letter (not an X) (BY trials).
Participants were instructed to always respond with a key “1” to the first letter presented (the cue). Depending on the second letter, the participant either had to respond with key “1” again (in case of AY, BX, BY), or participants had to respond with key “2” (in case of the AX trials) or do not respond at all (in case letter was followed by a number). Feedback was given when the participant was too slow, or made a wrong response.
In the version used for this experiment, no-go trials were used to reduce the utilisation of proactive control, which would take the BX trials off ceiling levels (Gonthier, Macnamara, Chow, Conway, & Braver, 2016). These trials (any letter followed by a number), which occurred unpredictably, would occur in about ~17% of the trials per block. The proactive behavioural index (PBI) was calculated for
accuracy as (BX – AY)/(AY + BX). Proactive control is reflected by a positive value, which is reflected by higher BX trial interference and lower AY interference. Reactive control is reflected by a negative value, which would indicate a higher AY trial interference and lower BX trial interference (Braver et al., 2009). The proactive control index for accuracy (PB acc), reaction time ( PB RT), and the Z- scores of RT (PB ZRT) was calculated for each participant. The task was implemented using E-prime software (Schneider, Eschman, & Zuccolotto, 2002).
Figure 1. The procedure of the AX-CPT. Example of the BX and the AY trials.
O-span
Working memory control (WMC) was measured using an automated version of the Operation-Span task (O-Span; Unsworth, Heitz, Schrock, and Engle, 2005). Participants had to remember a series of letters of varying lengths (3-7 letters), while performing a simple arithmetic operation (e.g. (3x2) + 4 = 11?)) (Turner & Engle, 1989). Participants’ response times were limited during the arithmetic
operation to make the rehearsal of letters more challenging. These response times were calculated by using the 2.5 SDs of personal average response time during the arithmetic practice set. The test procedure goes as follows: first, the arithmetic operation (duration mean time participant time plus 2.5 SD), then the letter (800ms), and at last, the recall screen (untimed).
Both the absolute score and the partial score were calculated for the O-span. The partial score reflects the number of letters recalled in the correct order, while the absolute score takes into account only the trial groups that were 100% correct before summing the number of letters recalled in the correct order (Conway, Kane, Bunting, Hambrick, Wilhelm, & Engle, 2005). The task was implemented using E-prime software (Schneider, Eschman, & Zuccolotto, 2002).
Materials and design
Stimuli for the experimental task were 160 three-to nine-letter nouns (median = 6) selected from the MRC psycholinguistic database (frequency range 1-10/million, Coltheart, 1981) and were visually presented on a computer monitor. The number of study-test cycles was determined for each age group on the basis of behavioural pilot tests, to ensure that age-related differences in neural activity were not due to the difficulty of retrieval. The performance was matched on Target hits made by younger and older adults (Dzulkifli, Herron, & Wilding, 2006; Herron & Rugg, 2003).
During the pilot study, it was determined that younger participants would receive four cycles, while the older participants would receive eight cycles. One cycle existed of a study phase and a test phase
(80 trials per cycle for young and 40 trials per cycle for old). For the younger adults, the study phase list comprised of 40 words, while the older participants list comprised of 20 words. These words were equally divided among the two semantic encoding tasks in two blocks (e.g. “Artist” and “Function” rating tasks). The same list length as the study was applied for the test phase (40 for young, 20 for old).
In the test phase, study words were randomly intermixed with no words repeated across cycles. Before each test word, preparatory cues were presented to participants to prepare them to remember whether they made an Artist or Function rating decisions on the test word. These preparatory cues switched predictably every two consecutive trials. Participants responded to 80 Targets and 80 Non-targets in total. A counterbalance for the response keys, the semantic encoding task block order (e.g. Artist task first or second), preparatory cue order, and Target/Non-Target test items was applied by making five different versions of the 160-word list. Each version had stimuli randomly assigned to a condition, after which it was randomly assigned to participants in both age groups. The task was implemented in MATLAB (version R2016b) with the COGENT 2000 toolbox (http://www.vislab.ucl.ac.uk/cogent_2000.php).
Procedure
The experimental task was a simplified version of the task used in Herron, et al., (2016), but with different encoding tasks (Artist and Function). Stimuli were presented in an Arial black font on a grey background, on a monitor 100cm away from the participant, which subtended 0.46° of horizontal and 5.2° vertical maximum visual angle. Participants started the experiment with several practice sessions to ensure the task was comprehended and to get familiar with the procedure and the keys of the task. The study phase began with an instruction screen specifying the encoding task that needed to be performed and the responses. In the Artist task in which the participant had to answer “How easy is it to draw the item?” indicating their judgement as ‘easy’ or ‘difficult’ or it could be the Function task “How many functions can you think of for the item?” indicating their judgement as ‘few’ or ‘many’. In both tasks, participants were asked to give their judgements verbally and then press spacebar to proceed to the next trial. Study trials were self-paced, with a maximum stimulus onset asynchrony (SOA) of 6100ms. Trials started with a black fixation cross for 300ms, then a red fixation cross for 300ms, followed by the word for 2000ms. After which a blank screen was presented either until the participant responded by pressing spacebar or until 3500ms (timeout). This was again preceded by a black fixation cross for 300ms indicating a new trial is to come, followed by a red fixation cross for 300ms. Before proceeding to the test phase, participants had to answer two question from a true or false questionnaire (pencil and paper, not on the screen) to prevent rehearsal of the study words (see appendix 1).
The test phase began with an instruction screen specifying the memory tasks and its appropriate responses. Explicit preparatory cues (e.g. Artist? Or Function?) indicating which task to remember were presented for 500ms, followed by a blank screen for 100ms, after which a black fixation cross was presented for 2000ms, which was followed by a blank screen for 100ms. Then the word stimulus was presented for 500ms, which was followed by a blank screen that would either disappear when the participant responded or after 3500ms (timeout). Participants had to respond ‘yes/no’ to the word stimulus depending on the preparatory cue given before the stimulus. The keys used were ‘Z’ and ‘M’, and these were counterbalanced across participants. After the response or the timeout, a jitter was applied which was either 400, or 500ms or 600ms. Then the end of the trial was announced with the red fixation cross for 1000ms, followed by a 100ms blank screen. To reduce the EEG
artefacts, participants were instructed to blink normally, but to minimise their horizontal eye movements by fixating on the centre of the screen.
Figure 2. Procedure for the study phase and test phase.
EEG Acquisition and pre-processing
Prior to the study and test phase, participants were fitted with an elastic cap housing 64 active silver/silver chloride electrodes using the extended International 10-20 system configuration (Jasper, 1958). CMS (Common Mode Sense) and DRL (Driven Right Leg) were two additional electrodes used to function as an online reference while concurrently supporting electrical noise rejection.
Furthermore, bipolar electrode pairs were placed on the outer canthi, and above and below the right eye to record horizontal and vertical electro-oculogram (EOG). Two more electrodes were placed on the left and right mastoid process. EEG was recorded at a 1024-Hz sampling rate with amplifier bandwidth of 0-208 Hz (3 dB) with the use of a 24-bit resolution BioSemi Active Two AD-box (http://www.biosemi.com/products.htm). EEG data were pre-processed in BrainVision Analyzer V2 (BVA; http://www.brainproducts.com/downloads.php?kid=9 ). The data were re-referenced offline using the signal average from the mastoid electrodes. This was then band-pass filtered (0.1- 80 Hz with 12 dB slope) and a 50 Hz notch filter was applied. The data were divided into epochs which were time locked to the 100 ms pre-onset of the preparatory cues and 2000 ms post-onset. Segments with excessive muscle artefact, drift, non-stereotyped EOG, or other artefacts were rejected following manual inspection. For the remaining stereotyped EOG, an independent component analysis (ICA) was applied to achieve ocular artefact correction by attenuating the EOG artefacts (Plank, 2013). Extremely noisy channels were interpolated using the spherical spline method (order of splines = 4; maximal degrees of polynomials = 10; lambda = 1*10-5). An additional low-pass filter of 20 Hz (12 dB slope) was used to smooth the epochs, after which a baseline correction was applied (-100ms to 0ms). The individual participant grand-average data were visually checked for residual artefacts.
Analysis strategy
Before the statistical test, the data were first tested on assumptions (e.g. normality, equal variances, and sphericity). To test the assumptions, the Shapiro-Wilk test was used for normality, the Levene’s test of equal variances for equal variances, and the Mauchly test of sphericity for sphericity. The conditions used to analyse the behavioural data were: Target hits (studied items designated as targets to which a target response was made), non-target correct rejection (CR; studied items designated as non-targets to which a non-target response was made). For the reaction times (RT), we examined median RTs. For the behavioural data the mixed factorial analysis of variance was used, separately on reaction times and accuracy, with factors Cue (Switch/Stay) x Task (Artist/Function) x ResType
(Hits/Non-target correct rejections) with between factors for Age Group (Young/Old).
To ensure that the age groups were comparable, the Mann-Whitney U test was used to measure the difference between groups on level of education. The Chi Square test was used for the analysis of difference between age and gender. For verbal IQ and Coding, the independent t-tests were used to
test for differences between the Age Group (Young/Old). The Mann-Whitney U test was used for the Verbal Paired Associates, the proactive control index, and the O-span absolute and partial scores, because the assumption of normality was violated. The behavioural data were analysed using SPSS 22.
For the ERP data, averages per switch/stay trials were calculated in BVA. These averages were calculated per participant separately for the following conditions: for artist task switch trials (e.g. from artist to function task), artist task stay trials (e.g. artist to artist task), function task switch trials (e.g. function to artist task), and function task stay trials (e.g. function to function). All these averages per condition per participant were added together and averaged to a cross-participant grand average per condition. After the grand averaging, the data were pooled per electrode group. The two
different pools used for the analysis were the left frontocentral channels (F1/F3/F5/FC1/FC3/FC5) and the right frontocentral channels (F2/F4/F6/FC2/FC4/FC6) as can be seen in Figure 2. The ERP data were analysed by a factorial mixed ANOVA with three factors, Cue (Switch/Stay) x Task
(Artist/Function) x Laterality (Left/Right) and between factor of Age Group (Young/Old). The mixed ANOVA was analysed using JASP (https://jasp-stats.org).
Figure 3. Electrode pools used. Pools (in circle) used for analysis in the factorial mixed ANOVA. Pools
are left lateral frontocentral channels: F1/F3/F5/FC1/FC3/FC5, and for the right lateral frontocentral pools F2/F4/F6/FC2/FC4/FC6.
Results
Behavioural and ERP data were analysed for all participants with sufficient artefact-free trials for ERP analysis (N= 26 young and N=19 old). The two age groups were not significantly different on
education level (U = 231, z = -0.43 p = 0.67). There was an equal amount of participants of both genders in the different age groups(Χ2(1) = 2.80 p = 0.094)
Standardised cognitive tests
Data from 27 younger adults, and 19 older adults were included. Except on the PC RT for young (n=26), n=16 for PC RT, PC ZRT for old. Table 1 summarises the results of the standardised tests. The TOPF measure of verbal IQ differed between Age Group (t(43) = -3.21, p = 0.003) as well as the Vocabulary test (not normally distributed; U = 159.5, Z = -2.017, p = 0.044). Older adults seemed to
have a higher verbal IQ and Vocabulary. Processing speed, also differed significantly between Age Group as expected (Coding task completion time: t(43) = 4.11, p < 0.001). The measures of the VPA were not normally distributed. Free recall is the only VPA measure not significant (U = 173.5, z = -1.714, p =0.087)
The two measures for WMC, absolute and partial, were not normally distributed. Both the O-span absolute (U = 146, z = -2.324, p =0.020) and the partial (U = 150, z = -2.232, p = 0.026) differed significantly between both age groups with younger adults seemingly scoring higher scores than the older adults.
The three measures of the AX-CPT, the PC acc, the PC RT, and the PC ZRT, were not normally
distributed. The PC acc (U = 153.5, z= -2.150, p = 0.032), the PC RT (U = 117, z= -2.659, p = 0.008), and the PC ZRT (U = 92, z= -3.390, p = 0.001) are all significantly different, which indicates a difference between the groups.
Table 1. Standardised cognitive test performance. Means of the performance of each cognitive test
is displayed in this table. One asterisk means that between-group comparisons were significant at p < 0.05, two asterisks means p < 0.025, three asterisks means p <0.001, and ns means non-significant.
Young Old Mean SD Mean SD Age 23.04 2.75 69.47 2.52 Education Level 3.54ns 0.91 3.32ns 1.25 Verbal IQ 110.62** 5.41 116.47** 6.80 Coding 89.35*** 12.09 72.217*** 15.89 Vocab 54.12* 5.48 57.68 * 7.08 VPA I Immediate recall 11.04*** 2.68 13.89*** 1.88 VPA II Delayed recall 11.92*** 1.57 13.79*** 1.40 Recognition 39.62 0.80 29.26*** 0.87 Free recall 12.12ns 1.95 13.32ns 2.58 O-Span Absolute 44.69** 16.13 30.42** 21.40 par. 59.31** 13.06 45.42** 22.76 PBI PC acc. -0.054* 0.18 -0.18 * 0.26 PC RT 0.083** 0.062 0.040 ** 0.26 PC ZRT 0.23** 5.19 -0.29** 1.08
Recognition memory accuracy
Table 2 summarises memory task performance. Mixed ANOVA on accuracy proportions with factors
of Cue (Switch/Stay), Task (Artist/Function), ResType (Target hits/ Non-Target CR) and Group revealed a three-way interaction between Task, ResType, and Age Group (F(1, 43) =10.498, MSE = 0.013, p =0.002, η2
p = 0.196), as well as a two-way interaction effect between Cue and Age Group (F(1,43) = 8.615, MSE = 0.009, p = 0.005, η2
p =0.167) and a two-way interaction effect between Cue and ResType (F(1,43) = 5.973, MSE = 0.009, p = 0.019, η2
p =0.122). However, there was no significant three-way interaction effect between Cue, Task, and Age Group (F(1,43) = 0.809, MSE = 0.008, p = 0.373), between Cue, Task, and ResType (F(1,43) = 0.540, MSE = 0.009, p = 0.466), and no four way interaction effect between Cue, Task, ResType, and Age Group (F(1,43) = 3.192, MSE = 0.009, p = 0.081).
Post hoc analysis showed a significant three-way interaction effect between Task, ResType, and Age Group (F(1,43) = 10.040, MSE = 0.007, p = 0.003, η2
p = 0.189). There is an interaction of Task with ResType in both the old (F(1,18) = 48.016, MSE = 0.023, P<0.001, η2
p =0.727) and the young (F(1,25) = 67.930, MSE = 0.006, p<0.001, η2
p = 0.731).
Older adults seemed to perform worse on non-target CR in the Artist task (M=0.672) than in the Function task (0.737). On the Target Hits, older adults performed worse on the Function Target Hits (M = 0.653) than on Artist Target Hits (M= 0.884).
Further post hoc analysis showed a significant group difference in the Artist task (ResType x Task,
F(1,43) = 74.983, MSE = 0.016, p < 0.001, η2
p = 0.636) and for the Function task (F(1,43) = 23.747, MSE = 0.016, p <0.001, η2
p = 0.021). There was a significant two-way interaction only in the Artist task between ResType and Age Group (F(1,43) = 10.093, MSE = 0.016, p = 0.003), not in the Function task (F(1,43) = 0.911, MSE = 0.016, p =0.345). Accuracy did differ according to retrieval task, in the Artist task the accuracy was higher for the Target Hits than Non-Target Correct Rejection. For the Function task, the participants performed worse on the Target Hits than on Non-Target Correct Rejection. For the interaction between Cue and Age, post hoc tests showed a significant effect for Cue for old (F(1,18) = 25.404, MSE = 0.007, p <0.001, η2
p = 0.585), but not for young (F(1,25) = 0.765, MSE = 0.010, p = 0.390). The older participants seemed to perform better on the stay trials (M=0.797) than switch (M=0.728), while younger participants perform around equal on both cues (switch, M =0.836; stay, M=0.848).
Table 2. Memory task performance. Mean accuracy proportions and RTs [ms] are shown for each
group separately by targeted task, trial type and cues. Mean target-non-target discrimination (D) for each task separately is also given (SD in brackets)
Target Artist Target Function
Target Non-Target Target Non-Target Switch Stay Switch Stay Switch Stay Switch Stay Youn g Accurac y 0.92(.08 4) 0.90(.11) 0.78 (.16) 0.83 (0.14) 0.79(.11) 0.79(0.15) 0.86(.1 1) 0.87(.091 ) D 0.72(.18) 0.66(.17) RT (ms) 1222(24 7) 1162 (237) 1505 (453) 1484(414) 1458(44 1) 1483(430) 1320 (318) 1415(406 ) Old Accurac y 0.89 (.097) 0.94(.094 ) 0.66(.1 7) 0.72(0.15) 0.65(.18) 0.68(0.19) 0.71(.1 3) 0.85(.13) D 0.68 (.19) 0.49(.25) RT (ms) 1616 (264) 1555(28 0) 2302 (395) 2204(39 0) 2391 (370) 2367(45 7) 2085 (403) 2025(35 3) Recognition memory RTs
Mixed ANOVA on reaction times with factors of Cue (Switch/Stay), Task (Artist/Function), ResType (Target hits/ non-target CR) and Group revealed a three-way interaction between Task ,Restype, and Age Group (F(1,43) = 23.34, MSE = 80589.77, p <0.001, η2
p = 0.352), as well as a two-way interaction between Cue and Age (F(1,43) = 5.917, MSE = 18703.09, p = 0.019,η2
p = 0.121).However, there was no significant interaction for Cue, Task, ResType and Age Group (F(1,43) = 0.050, MSE = 24164.92, p = 0.82), for Cue, Task, Age Group (F(1,43) = 0.777, MSE = 27816.88, p = 0.383), and for Cue, ResType, and Age Group (F(1,43) = 1.126, MSE = 40774.77, p = 0.295).
Post hoc analysis showed a significant interaction of Task and Restype for young (F(1,25) =28.611,
MSE = 74789.427, p<0.001, η2
p =0.534) and old (F(1,18) = 105.281, MSE = 88645.806, p<0.001, η2p = 0.854. Older adults seemed to perform slower on non-target CR in the Artist task (M = 2253.013) than in the Function task (M = 2055). On the Target Hits older adults performed worse on the Function Hits (M = 2378.842) than on Artist Hits (M=1585.697). The same pattern was found in the younger adults. Another post hoc analysis showed that there was a significant group difference in both Artist (F(1,43) = 28.359, MSE = 51491.433, p <0.001, η2
p = 0.397) and Function (F(1,43) = 7.694, MSE = 69510.510, p = 0.008, η2
p = 0.152) for ResType and Age Group. Younger participants are in general faster than the older participants. Reaction times to Target Hits or Non-Target Correct Rejections did differ between retrieval tasks. In the Artist Task, participants are quicker on the Target Hits, while in the Function task, participants are quicker on the Non-Target CR. Overall reaction times did not differ between younger and older adults by ResType (F(1,43) = 2.816, MSE = 40412.17, p = 0.101). For the interaction between Cue and Age, there seemed to be an interaction on cue for old (F(1,18) = 6.199, MSE = 23046.283, p = 0.023, η2
p = 0. 256). Older adults seemed to respond quicker on stay trials (M=2037), then on switch trials (M=2098). There was no interaction effect for the young on cue (F(1,25) = 0.313, MSE = 15575.996, p = 0.581, η2
p = 0.0012).
Preparatory ERPS
Primary analyses were conducted upon ERPs elicited by all types of trials, no matter the response. For all the ERP data, degrees of freedom and p-values were corrected for non-sphericity by using the Greenhouse-Geisser correction (Greenhouse & Geisser, 1959). Mean amplitudes of averaged ERPs were calculated for an a priori time window of 400-1200ms and 1200-2000ms based on by pilot data of the Herron Lab. The mean number of trials (ranges in parentheses) were calculated per Age Group. For the young, the mean number of trials were as follows: Artist Cues Switch = 29 (17-37), Artist Cues Stay = 28 (17-38), Function Cues Switch = 29 (18-39), and Function Cues Stay = 29 (20-36). For the old, the mean number of trials were as follows: Artist Cues Switch = 31 (22-39), Artist Cues Stay = 32 (24-40), Function Cues Switch = 30 (22-39), and Function Cues Stay = 30 (20-40). The ERPs in figure 4 are an average of the Switch/Stay trials per young and older adult.
In the 400-1200ms epoch, the only significant effect discovered was the one for Task (F(1,43) = 8.960, MSE = 4.061,p = 0.005, η2
p = 0.172). There were no significant interaction effects between Task, Cue, Laterality and Age Group (F(1.43) = 0.0.028, MSE =1.057, p = 0.868, η2
p = 0.001), as well as no significant main effect forCue (F(1,43) = 1.349, MSE = 5.86, p = 0.25, η2
p = 0.003), and Laterality (F(1,43) = 20.784, MSE = 0.422, p = 0.38, η2
p = 0.0018).
Figure 5. ERPS of the average of Switch and Stay trials per Age Group. ERPs elicited by preparatory
cues for each retrieval task from the left frontocentral pools, and the right frontocentral pools. On the left the weighted average of all the trials ERPs for the Stay are presented, on the right the weighted average of all the switch trials ERPs. The preparatory cue is presented at -600ms, and the retrieval cue
at 2100ms. The older adults are the top two left and right ERPs, the younger adults are the bottom two left and right ERPs. Red is the Function task, and blue is the Artist task.
For the 1200-2000ms, there were only main effects for Task (F(1,43) = 6.79, MSE = 6.31 p =0.013, η2 p = 0.136), Laterality (F(1,43) = 11.63, MSE = 1.82, p < 0.001, η2
p = 0.213). There was no significant main effect for Cue (F(1,43) = 0.078, MSE = 10.58, p = 0.781, η2
p = 0.002). Topographic maps as can be seen in figure 3, were made to determine what kind of lateralisation was happening.
Figure 6. Topography map for Switch and Stay trials per Age Group. The topographies maps from the
a priori analysis. Each map shows the scalp distribution of the effect obtained by subtracting the ERPs associated with the retrieval tasks from 400-1200ms. The scale bars to the right of the each map show the amplitude of the effect in microvolts.
However, no significant main effect for cue was discovered (F(1,43) = 0.21, p = 0.64) as well as no significant interaction effects between Task, Cue, Laterality and Age Group (F(1,43) = 0.0092, p = 0.92), between Task, Cue, and Laterality (F(1,43) = 0.018, p = 0.90). Exploratory analysis done on fifteen different EEG pools (AF7/F7/F5/F3, AFZ/F1/Fz/F2, AF8/F4/F6/F8, FC7/FC5/FC3, FC1/FC/FC2, FC4/FC6/FC8, T7/C5/C3, C1/Cz/C2, C4/C6/T8, TP7/CP5/CP3, CP1/CPz/CP2, CP4/CP6/TP8,
P7/P5/P3/PO7, P1/Pz/P2/POz, P4/P6/P8/PO8) in eight time windows from each 200 ms, starting from 400ms (400-600, 600-800, 800-1000, 1000-1200, 1200-1400,1400-1600, 1600-1800, 1800-2000). After a correction for multiple comparisons, there were no significant results in any of the time windows.
Discussion
The primary goal of this study was to investigate the effect of aging on the adoption and maintenance of retrieval orientation. The ERP data does not suggest an adoption and maintenance of retrieval orientations in both the younger and older participants. However, the behavioural data does seem to suggest that older adults use the preparatory cues differently from the younger adults. The
behavioural and ERP data did seem to differ between the results usually discovered on these task-switching paradigms such as by Herron and Wilding (2006) and Herron et al., (2016). There did seem to be a difference between the types of task as expected in the ERP data. The waveforms do suggest an age-effect for the preparatory cue. However, the behavioural data did suggest a difference in preparatory cue effect due to aging. Older adults were less accurate and less fast on switch trials than stay trials. There were more accurate responses on Target Hits in the Artist task and more accurate response on Non-Target Correct Rejections in the Function Task. Below, we discuss the implications of these and other findings for the age-related differences in preparatory cue processes.
The younger and older groups showed typical performance on the standardised cognitive tests. Both groups had a similar education level, but the verbal IQ and vocabulary were greater in the older group which correlates with the increase of knowledge with age (Li et al., 2004). The processing speed showed the age-related reduction as expected (Li et al., 2004). Older participants did not perform worse on the immediate recall and free recall as expected, which could be because of the lesser items the participants had to remember. The delayed recall and recognition were worse in
older adults as expected in Cullum et al., (1990). Older adults also did have a lower WMC than younger adults. Furthermore, it appeared that younger participants rely more on proactive control than older adults, because the PBI index is more positive going for younger participants. A more positive PBI index suggest more use of the proactive control.
As mentioned before, older adults were less accurate and less fast on switch trials than on stay trials. This is consistent with the task-switching literature (Monsell, 2003) and findings in Herron and Wilding (2006). This result also corresponds to the literature about switch cost (Arbuthnott & Frank, 2000). Switch cost reflects in a poorer performance for task switching trials than the same trials. However, younger adults performed around statistically equivalent for both cue types. This result can be explained by Rogers and Monsell (1995) in which they discovered that knowing when the tasks will switch will reduce the poorer performance. It could also mean that younger adults use more
proactive control. In the AX-CPT, proactive control is measured by the use of the cue before the probe. This system could be applied to the paradigm used in this study, because participants could use the cue to adopt a retrieval orientation. Older adults could have used reactive control more, which would make their performance on switch trials worse, but better on stay trials. The connection between proactive control and switch trials would be interesting to investigate further more
(Morcom, 2016).
The ERP data did not show a difference between the cue processing in older and younger adults. Even though there was no significant interaction effects in the ERP data, there does seem to be a possible age effect around the -400ms in the ERP waveform that seems more positive going(-400ms is 200ms after cue onset). In Herron et al., (2016), they did discover significant differences between retrieval tasks when frequently switched with an onset around 400ms and sustained effect till 1900ms at the end of the recording epoch. The same results were discovered in Herron and Wilding (2006), while in that paper the task paradigm was blocked and the effect onset was 400ms earlier than in the Herron et al., (2016) paper. Both papers did discover preparatory effects, which suggest that the problem of our study could be in the EEG data. The EEG data set used in this study was particularly noisy for participants, which could be an influence on the non-significant interaction effects we discovered. If the signal-to-noise (S/N) ratio is low, the amount of recording time to obtain a significant effect is increased (Luck, 2005). An increase in trials might help improve the S/N ratio and would perhaps result into a better results, as well as testing in an environment that is not humid or warm would improve the dataset and the results (Kappenman & Luck, 2010).
Another possibility for the non-significant interaction effects could be the difference in the difficulty in task. In Herron et al., (2016), the Target Hits were statistically equivalent in both the retrieval tasks. However, in this study there was a difference between the Target Hits and Non-Target Correct
Rejections in both tasks. In the Artist task, participants were faster and more accurate on the Target Hits, while in the Function task, participants were faster and more accurate on the Non-Target Correct Rejection. This seem to suggest that in the Function task, participants are more likely to be
correspond correctly when the retrieval cue was not associated with the retrieval goal (e.g incongruent, Artist cue with Function word). It could also be that older adults took more time to increase the likelihood of recovering this information accurately (Herron & Wilding, 2004). However, it could also indicate that participants have more difficulty with the Function task than the Artist task. The ERP data did not support the hypothesis, but the behavioural data did suggest an aging effect on the preparatory cue processes. It would be interesting to investigate this relationship further, because pilot data by the Herron lab showed promising results in younger adults. Furthermore, it would be interesting to apply the same technique as Gonthier et al., (2016) to the Artist/Function tasks. In Gonthier et al., (2016), strategies were applied to the AX-CPT to induce proactive control shifts. They used strategy training to induce more proactive control in the participants. In the strategy training, participants received the A cue followed by a X probe for 80% of the time. Participants were
instructed to mentally prepare for the most likely response to the probe during the cue interval. Duverne et al., (2009) did discover that the age effect on differential cue processing reflects the adoption of a specific retrieval strategy. It would be interesting to see what would happen if participant are taught a specific retrieval strategy to use.
In conclusion, our main findings show that behaviourally older adults seem to perform worse on switch trials than stay trials, which seems to correspond with older adults using more reactive control. However, these findings were not found in the ERP data, which could be because of the noisy data or the task difficulty. There does seem to be an aging effect on preparatory cue processes in specific to the methods how younger or older adults use it. Overall, the results look promising for an aging effect on the adoption of retrieval orientation during switch trials.
Biblography
Anderson, M. C., Bjork, R. A., & Bjork, E. L. (1994). Remembering can cause forgetting: retrieval dynamics in long-term memory. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 20(5), 1063.
Arbuthnott, K., & Frank, J. (2000). Executive control in set switching: Residual switch cost and task-set inhibition. Canadian Journal of Experimental Psychology/Revue canadienne de
psychologie expérimentale, 54(1), 33.
Braver, T. S. (2012). The variable nature of cognitive control: a dual mechanisms framework. Trends in cognitive sciences, 16(2), 106-113.
Braver, T. S., & Barch, D. M. (2002). A theory of cognitive control, aging cognition, and neuromodulation. Neuroscience & Biobehavioral Reviews, 26(7), 809-817.
Braver, T. S., Paxton, J. L., Locke, H. S., & Barch, D. M. (2009). Flexible neural mechanisms of cognitive control within human prefrontal cortex. Proceedings of the National Academy of
Sciences, 106(18), 7351-7356.
Braver, T. S., & West, R. (2008). Working memory, executive control, and aging. The handbook
of aging and cognition, 3, 311-372.
Bugg, J. M. (2014). Evidence for the sparing of reactive cognitive control with age. Psychology
and Aging, 29(1), 115.
Burgess, P. W. (1996). Confabulation and the control of recollection. Memory, 4(4), 359-412. Conway, A. R., Kane, M. J., Bunting, M. F., Hambrick, D. Z., Wilhelm, O., & Engle, R. W. (2005). Working memory span tasks: A methodological review and user’s guide. Psychonomic bulletin
Cullum, C. M., Butters, N., Tröster, A. I., & Salmon, D. P. (1990). Normal aging and forgetting rates on the Wechsler Memory Scale-Revised. Archives of Clinical Neuropsychology, 5(1), 23-30.
Craik, F. I. M. (1984). Age differences in remembering. Neuropsychology of memory, 3-12.
Craik, F. I., & McDowd, J. M. (1987). Age differences in recall and recognition. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 13(3), 474.
Dew, I. T., Buchler, N., Dobbins, I. G., & Cabeza, R. (2011). Where is ELSA? The early to late shift in aging. Cerebral cortex, 22(11), 2542-2553.
Duverne, S., Motamedinia, S., & Rugg, M. D. (2009). Effects of age on the neural correlates of retrieval cue processing are modulated by task demands. Journal of Cognitive
Neuroscience, 21(1), 1-17.
Dzulkifli, M. A., Herron, J. E., & Wilding, E. L. (2006). Memory retrieval processing: Neural indices of processes supporting episodic retrieval. Neuropsychologia, 44(7), 1120-1130. Dzulkifli, M. A., & Wilding, E. L. (2005). Electrophysiological indices of strategic episodic retrieval processing. Neuropsychologia, 43(8), 1152-1162.
Elward, R. L., & Wilding, E. L. (2010). Working memory capacity is related to variations in the magnitude of an electrophysiological marker of recollection. Brain research, 1342, 55-62. Engle, R. W., & Kane, M. J. (2003). Executive attention, working memory capacity, and a two-factor theory of cognitive control. Psychology of learning and motivation, 44, 145-199. Friedman, D., & Johnson, R. (2000). Event related potential (ERP) studies of memory ‐ encoding and retrieval: a selective review. Microscopy research and technique, 51(1), 6-28. Gonthier, C., Macnamara, B. N., Chow, M., Conway, A. R., & Braver, T. S. (2016). Inducing proactive control shifts in the AX-CPT. Frontiers in psychology, 7.
Greenhouse, S. W., & Geisser, S. (1959). On methods in the analysis of profile data. Psychometrika, 24(2), 95-112.
Herron, J. E., & Rugg, M. D. (2003). Retrieval orientation and the control of recollection. Journal of Cognitive Neuroscience, 15(6), 843-854.
Herron, J. E., Evans, L. H., & Wilding, E. L. (2016). Electrophysiological evidence for flexible goal-directed cue processing during episodic retrieval. NeuroImage, 132, 24-31.
Herron, J. E., & Wilding, E. L. (2004). An electrophysiological dissociation of retrieval mode and retrieval orientation. Neuroimage, 22(4), 1554-1562.
Herron, J. E., & Wilding, E. L. (2006). Neural correlates of control processes engaged before and during recovery of information from episodic memory. Neuroimage, 30(2), 634-644.
Jasper, H. H. (1958). The ten twenty electrode system of the international federation.
Jacoby, L. L., Shimizu, Y., Daniels, K. A., & Rhodes, M. G. (2005b). Modes of cognitive control in recognition and source memory: Depth of retrieval. Psychonomic bulletin & review, 12(5), 852-857.
Jacoby, L. L., Shimizu, Y., Velanova, K., & Rhodes, M. G. (2005a). Age differences in depth of retrieval: Memory for foils. Journal of Memory and Language, 52(4), 493-504.
Johnson, M. K., Kounios, J., & Nolde, S. F. (1997). Electrophysiological brain activity and memory source monitoring. NeuroReport, 8(5), 1317-1320.
Kane, M. J., & Engle, R. W. (2003). Working-memory capacity and the control of attention: the contributions of goal neglect, response competition, and task set to Stroop
interference. Journal of experimental psychology: General, 132(1), 47.
Kappenman, E. S., & Luck, S. J. (2010). The effects of electrode impedance on data quality and statistical significance in ERP recordings. Psychophysiology, 47(5), 888-904.
Keating, J., Affleck-Brodie, C., Wiegand, R., & Morcom, A. M. (2017). Aging, working memory capacity and the proactive control of recollection: An event-related potential study. PloS
one, 12(7), e0180367.
Koen, J. D., & Yonelinas, A. P. (2014). The effects of healthy aging, amnestic mild cognitive impairment, and Alzheimer’s disease on recollection and familiarity: a meta-analytic review. Neuropsychology review, 24(3), 332-354.
Li, S. C., Lindenberger, U., Hommel, B., Aschersleben, G., Prinz, W., & Baltes, P. B. (2004). Transformations in the couplings among intellectual abilities and constituent cognitive processes across the life span. Psychological Science, 15(3), 155-163.
Li, J., Morcom, A. M., & Rugg, M. D. (2004). The effects of age on the neural correlates of successful episodic retrieval: an ERP study. Cognitive, Affective, & Behavioral
Neuroscience, 4(3), 279-293.
Light, L. L. (1991). Memory and aging: Four hypotheses in search of data. Annual review of
psychology, 42(1), 333-376.
Luck, S.J. (2005). An Introduction to the Event-Related Potential Technique. Cambridge, MA: MIT Press;
Mecklinger, A. (2010). The control of long-term memory: brain systems and cognitive processes. Neuroscience & Biobehavioral Reviews, 34(7), 1055-1065.
Morcom, A. M. (2016). Mind over memory: Cuing the aging brain. Current Directions in
Psychological Science, 25(3), 143-150.
Morcom, A. M., & Rugg, M. D. (2004). Effects of age on retrieval cue processing as revealed by ERPs. Neuropsychologia, 42(11), 1525-1542.
Old, S. R., & Naveh-Benjamin, M. (2008). Differential effects of age on item and associative measures of memory: a meta-analysis.
Richmond, L. L., Redick, T. S., & Braver, T. S. (2015). Remembering to prepare: The benefits (and costs) of high working memory capacity. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 41(6), 1764.
Rogers, R. D., & Monsell, S. (1995). Costs of a predictible switch between simple cognitive tasks. Journal of experimental psychology: General, 124(2), 207.
Rugg, M. D., & Curran, T. (2007). Event-related potentials and recognition memory. Trends in
cognitive sciences, 11(6), 251-257.
Rugg, M. D., & Morcom, A. M. (2005). The relationship between brain activity, cognitive performance and aging: the case of memory. Cognitive neuroscience of aging: Linking
cognitive and cerebral aging, 132-154.
Rugg, M. D., & Wilding, E. L. (2000). Retrieval processing and episodic memory. Trends in
cognitive sciences, 4(3), 108-115.
Salthouse, T. A. (2000). Aging and measures of processing speed. Biological psychology, 54(1), 35-54.
Schneider, W., Eschman, A., & Zuccolotto, A. (2002). E-Prime reference guide. Psychology Software Tools, Incorporated.
Spencer, W. D., & Raz, N. (1995). Differential effects of aging on memory for content and context: a meta-analysis.
Tulving, E. (1983). Elements of episodic memory. Oxford, UK: Clarendon.
Tulving, E. (2002). Episodic memory: from mind to brain. Annual review of psychology, 53(1), 1-25
Turner, M. L., & Engle, R. W. (1989). Is working memory capacity task dependent?. Journal of
memory and language, 28(2), 127-154.
Unsworth, N. (2007). Individual differences in working memory capacity and episodic
retrieval: examining the dynamics of delayed and continuous distractor free recall. Journal of
Unsworth, N., Heitz, R. P., Schrock, J. C., & Engle, R. W. (2005). An automated version of the operation span task. Behavior research methods, 37(3), 498-505.
Plank, M. (2013). Downloads & Support Ocular Correction ICA. Brain, 49.
Wechsler, D. (1997). WAIS-III: Wechsler adult intelligence scale. Psychological Corporation. Wechsler, D. (2008). Wechsler adult intelligence scale–Fourth Edition (WAIS–IV). San Antonio,
TX: NCS Pearson, 22, 498.
Wechsler, D. (2011). Test of premorbid functioning. UK version (TOPF UK). UK: Pearson
Corporation.
World Health Organisation (2015). World report on aging and health. Retrieved from http://apps.who.int/iris/bitstream/10665/186468/1/WHO_FWC_ALC_15.01_eng.pdf?ua=1
Yonelinas, A. P. (2002). The nature of recollection and familiarity: A review of 30 years of research. Journal of memory and language, 46(3), 441-517.
Appendix 1
