The Effects of Ageing on the ERP Correlates of Recognition

To date, ERP research on the influence of ageing on the neural correlates of episodic retrieval has been somewhat limited. The current section will provide a review of this

limited literature, and, because the focus has primarily been on the effects of ageing on left parietal and right frontal effects, these modulations will be examined in particular.

4.3.1 Continuous Recognition Studies

The most common task employed during early ERP investigations of recognition memory and ageing was the continuous recognition paradigm. In this paradigm participants were presented with a series of stimuli, some of which were repeated after different lags (delays), and participants had to differentiate between previously seen and new items (Rugg et al., 1997; Swick and Knight, 1997). Behavioural results indicate that older adults show reduced performance on continuous recognition tasks compared to the young, especially as the delay increases. ERP findings showed that a centro-parietal positive old/new effect (discussed below), significant for all lag conditions in the young group, was reduced in the older group at short lags and non-significant at longer lags. These studies employed visually presented words as stimuli, however, continuous recognition studies using auditory stimuli (Minamoto et al., 2001) and pictorial stimuli (Nielson-Bohlman and Knight, 1995) show that the age-related ERP changes are generalisable beyond visual words. Importantly, Nielson-Bohlam and Knight (1995) noted age equivalence in the ERPs to new items, suggesting that the ERP age difference in the old/new effects indicates changes in retrieval processes.

The centro-parietal effect evident in continuous recognition tasks is more bilaterally distributed than the left parietal correlate of recollection, but they both occur around the same time (approximately 400-900ms post stimulus). The two components are assumed to be related, with the topographic difference simply reflecting design differences between continuous recognition tasks and study/test paradigms (e.g. Friedman, 2000).

However, because it is open to debate whether or not the delay even in the long lag

condition is great enough to test long-term memory, the centro-parietal modulation may not index long-term memory processes.

It is therefore questionable whether continuous recognition studies provide evidence of age-related changes in episodic memory. The study/test paradigm overcomes the long-term memory criticism, where a longer delay occurs between study and test than the long lag conditions of continuous recognition tasks. Using a study/test paradigm, Morcom and Rugg (2004) found a left parietal effect in young adults for the recognition of words following the encoding of either pictorial or verbal stimuli. This effect was absent in older adults for both pictorial and verbally encoded items, even when performance was age matched.

In line with dual process theory, the absence of the left parietal effect in the Morcom and Rugg study indicates that the elderly were relying less on recollection based remembering than the young. This age difference can be explained by either an ageing deficit in recollection or an under use of recollection when the task can be completed based on familiarity. To address these possibilities, the following section will present the findings regarding age-related changes in the neural correlates of episodic retrieval using a source memory task, which promotes recollection-based retrieval.

4.3.2 Source Memory Studies

During a typical source memory task, items are presented under two study conditions (e.g. different voices or lists) and, at test, participants are required to remember the source by making one-, two-, or three-stage judgements. For example, in a study phase Trott et al. (1997) presented two lists of sentences to young and older adults with the requirement to learn the nouns from each sentence, along with the list in which they

were studied. During the test phase, participants were instructed to discriminate between old and new nouns and then, for words judged old, to make a remember/know judgement followed by a source decision. Behaviourally, the elderly had a source memory deficit compared to the young, but no between group difference was present in the rate of remember responses. Electrophysiologically, both correct source judgements and remember responses were associated with a left parietal effect of equivalent

magnitude in both age groups. In addition, the right frontal effect was present in the young but was significantly reduced in the elderly. Wegesin et al. (2002) noted comparable findings when they replicated the aforementioned study, including a variation to improve the older adults’ near chance performance.

However, other source memory studies are inconsistent with the foregoing findings.

Mark and Rugg (1998) required participants to remember the source of words heard in a male or female voice, whereas Li et al. (2004) asked participants to recollect which of two encoding tasks they had performed on pictorial stimuli. Source accuracy did not differ between age groups. Both of these studies found no right frontal differences between age groups, however, they did observe an age-related reduction in the left parietal effect.

A negative left/central old/new effect has also been reported in older adults, maximal at around 1100ms, in three of the previous source memory studies (Trott et al., 1997;

Wegesin et al., 2002; Li et al., 2004), and even evident when performance was equated across age groups (Li et al., 2004). Wegesin et al. (2002) compared this negativity to the late posterior negative slow wave (LPN) recorded in young adults, surmising that the more posterior distribution in the young reflected attenuation at central electrodes by the right frontal effect. This interpretation is tentative, however, as Li et al., (2004) showed that when the right frontal effect was equivalent across age groups, the

negativity in the elderly was still more centrally distributed. Two alternative accounts of the functional significance of the modulation seen in the elderly are: Wegesin et al.

(2002) incorporated a two stage response in their source memory task (old/new then source) and noted that the modulation occurred after the mean RT response for the initial old/new decision. Accordingly, they proposed that the modulation represents the engagement of different/compensatory processes associated with the search for, or retrieval of, source information. Alternatively, following the use of pictorial stimuli, Li et al. (2004) suggested that it reflects older adults greater need to reconstruct perceptual details of the study context to make source decisions. It remains unclear which of these interpretations, if any, reflects the functional significance of the left/central negativities.

Consistent with the notion that the elderly engage compensatory processes to reduce age-related source memory deficits, PET studies have shown that, in certain

circumstances, older adults show bilateral prefrontal cortex activation in comparison to right sided only in the young (Backman et al., 1997; Madden et al., 1999; Cabeza et al., 2002). In particular, in a PET study, Cabeza et al (2002) grouped older adults according to their performance on neuropsychological memory tests. During a source recognition task, high performers on the memory tests showed bilateral prefrontal cortex activation, compared to right prefrontal activation only in the low performers.

4.3.3 Ageing and the Bilateral Frontal Effect

To date, very few studies have reported the effects of ageing on early onsetting ERP effects. Collectively, the studies that have reported early frontal modulations (evident between 300 and 600ms post stimulus) showed similar magnitudes across age groups, but a bilateral distribution for the young and a right sided distribution for the elderly (Wegesin et al., 2002; Morcom and Rugg, 2004). These distributional differences,

while difficult to interpret, may reflect the age-related engagement of different cognitive processes, or ageing changes in brain morphology that have altered the alignment of the neural generators of the bilateral frontal effect (Rugg and Morcom, 2004).