~The hippocampus~

 My research combines behavioural, neuropsychological and structural and functional neuroimaging approaches to understanding the cognitive and neural bases of autobiographical memory (AM): memory for both personally experienced events (personal episodic memory) and facts about oneself (personal semantic memory).  I have a particular interest in the neural networks mediating the retrieval of autobiographical memory (AM), how these cognitive processes and neural networks change with dysfunction in various populations (e.g., healthy aging, temporal lobe epilepsy, dementia), and how such changes may impact upon the functioning of self and identity, and the ability to simulate future events.

Additionally, I am interested in the role of the hippocampus in memory as an integrator of information.  Thus, some of my focus in studying AM is to look at the integration of different recollective information upon retrieval, such as contextual information, visual imagery and emotional information.  I am also interested in the way in which the hippocampus acts to bind information upon encoding, and have examined how this changes with healthy aging. The integration of information is not only important to memory encoding and retrieval, but it appears to be important when simulating or imagining future events. My current research program is examining this possibility.

Within cognitive neuroscience, functional neuroimaging is a predominant research methodology.  Much of my research at Harvard University and the University of Toronto has involved the use of functional magnetic resonance imaging. As such, I am very interesting in using different statistical approaches to the analyses of imaging data, including parametric modulation and functional and effective connectivity analyses.  

The Role of the Hippocampus in Autobiographical Memory

Recent neuroimaging studies report preferential hippocampal engagement during AM retrieval.  I am interested in the attributes of AM which give rise to this preferential hippocampal activation.  Using fMRI, I investigated whether this was due to the temporal specificity, recency or recollective qualities of AMs, such as detail, emotionality and personal significance (Addis et al., 2004, Hippocampus). I found no effect of temporal specificity on the level of hippocampal activation.  Activation of the left hippocampus during specific AM retrieval did vary with the level of detail, personal significance, and at a subthreshold level, emotionality, when the effect of recency was covaried out. Further, during general AM retrieval, all three recollective qualities modulated activity in the right hippocampus. Although the recency of specific AMs modulated hippocampal activation bilaterally, this effect dissipated in the left hippocampus when detail or emotionality was included as a covariate, and was no longer present in either hippocampus when personal significance was taken into account. These results suggest that recollective qualities are important predictors of hippocampal engagement during AM retrieval independent of factors such as recency.

Neural Networks Supporting Autobiographical Memory

I have also examined the networks supporting retrieval of specific AMs (i.e., AMs of unique events specific in time and place, e.g., breaking my leg) and general AMs (i.e., memories of events which occur repeatedly, e.g., commuting to grad school; Addis et al., 2004, NeuroImage ). Using functional connectivity analyses ( spatiotemporal partial least squares), I identified distinct sub-networks supporting these types of AMs. Specific AMs were associated with posterior cortical activity supporting visuospatial and imagery processes, while general AMs engaged regions involved in conceptual retrieval (e.g., temporal pole). These networks differed temporally, in line with previous cognitive studies, such that activity associated with general AMs peaked 2 seconds before that associated with specific AMs. Despite these differences, I found a core network of regions functionally connected with the hippocampus that was engaged by both types of AM.

In another study, I examined the networks mediating the retrieval of AMs varying in levels of recollective qualities (i.e., detail, emotionality and personal significance) using effective connectivity analyses (Addis, McIntosh et al., in prep). This also revealed a core network, including the hippocampus, common to the retrieval of all AMs types. The retrieval of different recollective qualities again was associated with recruitment of additional regions into the common network, specifically regions involved in the processing of information related to detail, emotionality and personal significance. For example, the anterior cingulate (affective division) was recruited into the network during emotional AM retrieval, and further, negative connectivity between prefrontal and hippocampal regions was unique to this AM type, possibly reflecting an initial suppression of emotional memory retrieval (Conway et al., 2001).

The second major goal of this research was to examine how AM retrieval and the associated neural network changes with left hippocampal damage in temporal lobe epilepsy patients (Addis, Moscovitch et al., submitted). Firstly, I confirmed the presence of significant left hippocampal atrophy using a linear measure of hippocampal volume (MTL width; Gao et al., 2004). Secondly, I used the Autobiographical Interview (Levine et al., 2002) to probe the integrity of AM and found mild deficits in the episodic but not the semantic aspects of AM. During the retrieval of residual AMs, patients exhibited significant reductions in activation of the hippocampus, as well as other regions in the AM network. The effective connectivity of the AM network was dramatically different between patients and controls, with an apparent “bypassing” of the left hippocampus. Together, the findings from this body of research suggest the hippocampus is a key structure supporting AM retrieval, constituting the “hub” of the AM network. When damaged, the network fails to engage normally and the ability to vividly recollect the past is diminished

remembering the past AND imagining the future: The CONSTRUCTIVE EPISODIC SIMULATION HYPOTHESIS

My most recent line of research seeks to understand how AMs of past events are used in the simulation of future events. The ability to shift our perspective from the present to the imagining of future scenarios is critical to both psychological and social functioning. In a recent integration of the simulation and memory distortion literatures, Daniel Schacter and I proposed the constructive episodic simulation hypothesis (Schacter and Addis, 2007, Nature; Schacter and Addis, 2007, Phil Trans Royal Soc B ) in which we outline how the constructive nature of episodic memory is well-suited, and perhaps adapted specifically for, the simulation of future events. Although the constructive character of episodic memory means it is prone to various kinds of errors and distortions, it is this attribute of the system which enables individuals to imagine future scenarios. Because future events are not exact replicas of the past, the simulation of future episodes requires a system that can flexibly extract and re-combine elements of previous experiences.

Consistent with the constructive episodic simulation hypothesis, I recently demonstrated considerable overlap in the neural networks engaged by past and future events, both during the construction and the subsequent elaboration of these events (Addis et al., 2007, Neuropsychologia). Specifically, this fMRI study showed left hippocampus and posterior visuospatial regions to be commonly engaged by past and future event construction. However, neural differences were also observed during this phase, with future events recruiting regions involved in prospection (right frontopolar cortex), generative processing (left inferior frontal cortex), and notably, relational processing (right hippocampus). In contrast, elaboration was marked by striking neural overlap between past and future events, such that the AM network supporting past event retrieval was fully recruited by future events.

Although the elaboration of past and future events recruits a common neural network, regions within this network may respond differentially to the phenomenological characteristics of events (e.g., level of detail) depending on whether the event is in the past or future (Addis et al., 2007, CNS abstract ). Using parametric modulation analyses, I found two effects that were unique to future events: (1) increasing temporal distance of future events (from the present) was associated with increasing hippocampal activity, possibly reflecting a higher relational load when recombining increasingly disparate details, and (2) increasing levels of detail strongly modulated bilateral hippocampal activity. These findings highlight hippocampal involvement in future event simulation. In a current fMRI study, I am attempting to further characterize the nature of this involvement, by probing the novelty of future events. Given that novelty is another factor that can influence hippocampal activity, this study will be important in better understanding hippocampal contributions to future events, particularly as these events are novel by definition (i.e., they have not yet occurred).

I have also examined how episodic simulation changes with healthy aging (Addis et al., submitted). Using an adapted version of the Autobiographical Interview (Levine et al., 2002), I found that age-related reductions in the episodic specificity of past events extends to future events. Further, the ability to generate episodic details for past and future events was strongly correlated. Notably, this ability was also correlated with relational memory performance but not executive functioning. These findings are consistent with the constructive episodic simulation hypothesis and the idea that relational processing is an important component of episodic simulation. Currently, I am examining the episodic specificity of past and future events, relational memory performance and prospective memory abilities in patients with Alzheimer's disease and frontal lobe lesions, with the aim of further characterising the cognitive processes critical to episodic simulation and the projection of self into the future.

The Role of the Hippocampus in relational processing

A major role of the hippocampus during AM retrieval appears to be the integration of different recollective aspects of a memory, highlighting the relational processing functions of this structure. I have recently expanded my research to encompass relational memory, specifically the encoding of relational information. In these studies, I used a semantic-relatedness paradigm, in which triads with varying numbers of semantic relations were presented during encoding. Triads with fewer associations had higher generative load, whereas triads with more associations had higher relational load. I found that hippocampal activity was modulated by the relational load of the encoding task while the left inferior frontal activity was modulated by generative load (Addis and McAndrews, 2006). Further, there was evidence of strong positive connectivity between these regions which did not differ according to the number of associations provided

In my post-doctoral work, I have used the same paradigm to examine whether age-related declines in relational encoding reflect dysfunction of inferior frontal gyrus linked with deficient generation of associations, and/or hippocampal dysfunction linked with impoverished binding of associations (Addis, Giovanello and Schacter, in prep) . As expected, older adults exhibited decreased levels of activity in both regions relative to young adults. Further, generative load did not modulate the inferior frontal gyrus. However, the hippocampus responded significantly to relational load, suggesting that when provided with associations to bind, hippocampal activity in older adults is comparable to young, consistent with increased recognition accuracy under such conditions. Importantly, this finding suggests that age-related impairments in relational encoding are likely the result of deficits in prefrontal engagement and the generation of associations, rather than the binding of associations by the hippocampus.

The contributions of autobiographical memory to the self

Another strand of my ongoing research considers the role that AM plays in identity. By integrating methods from social psychology and neuropsychology, I examined the impact of AM loss on the integrity of identity (measured with the Twenty Statements Test and the Tennessee Self Concept Scale) in patients with Alzheimer's disease (Addis and Tippett, 2004, Memory ). This revealed that loss of late childhood/early adulthood AMs had the most significant effects on identity. For instance, with AM loss identity became increasingly abstract in nature, even when controlling for general cognitive decline. I am currently working on a two-year follow-up of a subset of patients and controls to explore the progression of AM and identity deficits as Alzheimer's disease advances. I am also interested in how AM contributes to specific aspects of identity. I recently authored a chapter examining how episodic and semantic AM contribute to the content and continuity of identity (Addis and Tippett, in press). I also proposed a distinction between narrative and phenomenological forms of self-continuity, the latter being based primarily on episodic memory and the ability to project the self over time.