Memory problems are amongst the most common complaints as we grow older. These problems relate mainly to difficulties encoding new, lasting memories (rather than retrieving older memories). Such 'anterograde, long-term memory' is enabled by multiple cognitive processes, supported by distinct neural networks, with differential life-span trajectories. For example, 'source memory' - the ability to recall the context in which an item was encountered - involves specialised regions within the Medial Temporal Lobe (MTL) such as the hippocampus, and tends to show greatest impairment in old age. 'Item memory'on the other hand - the ability to recognise individual items - is thought to be supported primarily by the perirhinal region of the MTL, and also shows impairment in old age. Item and source memory are examples of conscious or 'explicit' memory. A third type of memory is 'priming', an example of 'implicit' memory that occurs without awareness of a past encounter. Priming is believed to reflect plasticity in neocortical regions outside the MTL, such as occipitotemporal cortex in the case of the visual object priming, and is much less affected by age. The slower rate of neural loss within such sensory regions is generally consistent with these findings.
However it is also believed that neural reorganisation can compensate for this neural change, for example, by a greater degree of bilateral prefrontal cortex (PFC) activity in older people during encoding and retrieval of memories. Furthermore, it is well-established that memory is generally superior for emotional relative to neutral stimuli, which is often attributed to modulation of MTL by the amygdala. This emotional memory advantage also changes with age, depending on the positive or negative valence of the emotions induced. Finally, while most memory studies have focused on old age, it is likely that brain systems supporting memory begin to decline earlier than typically thought, but this is masked by compensatory cognitive processes, such as more effortful encoding, perhaps supported by PFC. Given that the PFC is still maturing in the third decade of life, differences between various types of memory, and their emotional modulation, are likely to be seen in early, as well as late, years of life.
We will address the above hypotheses about priming, item memory and source memory for stimuli with positive, negative and neutral associations. We will test how each type of memory is affected by age, by emotional valence, and by the interaction between these factors. We will also test whether age/valence effects are modulated by each participant?s scores on other tests of attention and basic emotional responses (see Attention and Emotion Domains). We will then examine where any selective memory effects correlate with MRI measures of grey-matter volume and white-matter connectivity, and with functional activity as recorded by fMRI and MEG+EEG, with a focus on encoding-related changes in functional connectivity that predict subsequent memory.
Current behavioural investigations examine the effect of age on item memory, source memory, and priming. Preliminary results show an expected age-related decline in both item and source memory (figure below, left and middle). However, despite age-related declines in memory, our initial findings also support previous findings by demonstrating preserved priming across the lifespan (figure below, right):
Memory, valence and intelligence
The memory task is a modified CID-R paradigm (Berry et al, 2012) that provides trial-by-trial measures of priming, familiarity and recollection. Previous work has suggested that ageing has dramatic effects on recollection, but little to no effect on priming. Furthermore, the emotional valence of information has been shown to interact with age, whereby the memory-advantage for negative information in the young shifts to an advantage for positive information in older people (“positivity bias”). The valence of stimuli was therefore manipulated, to relate memory to other CamCAN data on emotional-processing. At study, neutral objects are paired with background scenes that are either of positive, neutral and negative valence (from the IAPS). During each test trial, only a studied or unstudied object is presented, which is first identified (named) in the presence of pixelwise noise (indexing priming); second, recognised from the study phase (after noise removed), using a 4-level confidence rating (from which familiarity can be derived), and third, used to cue recall of the associated scene at study (indexing recollection).
These asymmetrical age effects will be followed up with an examination of the role of regional structural integrity. For example, one explanation for these differential effects of age on different types of memory is that the MTL is more affected by ageing than the occipitotemporal regions associated with visual object priming. Future analyses will also examine the role of emotional processing on memory and the effects of age. A planned fMRI study will expand on structural analyses to examine the potential that prefrontal cortex can play a compensatory role in preserving performance.
Preliminary data from the first 107 CC700 volunteers (Fig bottom,left) confirms a large effect of age on recollection (R=-.68), a smaller effect on recognition (R=-.54) but no significant effect on priming (R=-.12; despite a strong effect of age on overall identification accuracy, R=-.59). This supports previous claims that age differentially affects these different types of memory. Importantly, the effect of age on recollection remained after partialing-out the effect of age on IQ (as indexed by Cattell), suggesting that the memory impairment does not reflect general intellectual decline (Fig bottom, middle). Emotional valence exerted a clear effect on recollection, with greater recall of negative than positive than neutral scenes, and a trend for a less of an age-decline in recall of positive than negative scenes, p=.06, two-tailed (Fig bottom, right).