Cognition and dementia in older patients with epilepsy

Abstract

With advances in healthcare and an ageing population, the number of older adults with epilepsy is set to rise substantially across the world. In developed countries the highest incidence of epilepsy is already in people over 65 and, as life expectancy increases, individuals who developed epilepsy at a young age are also living longer. Recent findings show that older persons with epilepsy are more likely to suffer from cognitive dysfunction and that there might be an important bidirectional relationship between epilepsy and dementia. Thus some people with epilepsy may be at a higher risk of developing dementia, while individuals with some forms of dementia, particularly Alzheimer's disease and vascular dementia, are at significantly higher risk of developing epilepsy. Consistent with this emerging view, epidemiological findings reveal that people with epilepsy and individuals with Alzheimer's disease share common risk factors. Recent studies in Alzheimer's disease and late-onset epilepsy also suggest common pathological links mediated by underlying vascular changes and/or tau pathology. Meanwhile electrophysiological and neuroimaging investigations in epilepsy, Alzheimer's disease, and vascular dementia have focused interest on network level dysfunction, which might be important in mediating cognitive dysfunction across all three of these conditions. In this review we consider whether seizures promote dementia, whether dementia causes seizures, or if common underlying pathophysiological mechanisms cause both. We examine the evidence that cognitive impairment is associated with epilepsy in older people (aged over 65) and the prognosis for patients with epilepsy developing dementia, with a specific emphasis on common mechanisms that might underlie the cognitive deficits observed in epilepsy and Alzheimer's disease. Our analyses suggest that there is considerable intersection between epilepsy, Alzheimer's disease and cerebrovascular disease raising the possibility that better understanding of shared mechanisms in these conditions might help to ameliorate not just seizures, but also epileptogenesis and cognitive dysfunction.

Figure 1

Trajectories of cognitive decline with ageing. The schematic illustrates how cognitive function in people with epilepsy might decline compared to healthy ageing (blue). One model proposes that an initial brain insult (‘initial hit’) leads to cognitive decline in epilepsy patients simply running parallel to but below the normal trajectory (dashed yellow). These individuals start from lower cognitive performance and also reach the threshold for functional impairment or dementia earlier. An alternative model is that while an initial hit might be a neurodevelopmental disorder or traumatic brain injury, subsequent development of epilepsy is in effect a ‘second hit’, which leads to further deviation from the normal trajectory (solid yellow). A third proposal is that, with increasing time, the trajectory of cognitive decline in people with epilepsy deviates further from that in healthy individuals leading to accelerated cognitive ageing (red). Inspired by a figure used by Breuer et al. (2016).

Figure 2

The intersections of Alzheimer’s disease, epilepsy and vascular disease. Several overlapping pathologies (right) can contribute to development of late-onset epilepsy as well as the development of dementia. In particular, vascular risk factors (left) are common in people with epilepsy. These may represent modifiable risk factors for both the development of dementia and of epileptogenesis.

Figure 3

Mouse model of Alzheimer’s disease shows abnormal spiking. The human amyloid precursor (hAPP) transgenic mouse of model of Alzheimer’s disease shows deficits of spatial memory on the Morris water maze task, taking longer to find the platform (top left) and with abnormal spiking activity over left and right temporal and parietal cortex (bottom left). Reduction of tau by creating a hAPP mouse, which does not produce tau (hAPP/Tau^-/-) normalizes both spatial memory and the EEG (top and bottom right, respectively), despite the fact that both models have similar amyloid plaque deposition (middle). Modified with permission from Roberson et al. (2011).

Figure 4

Hippocampal interictal epileptiform discharges couple with frontal activity. Rodent studies: (A) Medial prefrontal cortical (mPFC; red) and hippocampal (blue) local field potentials (LFPs) recorded during non-REM sleep in a kindling model of TLE. This is an example of a putative hippocampal IED-evoked spindle (oscillation) in frontal cortex. Similar findings occurred in REM sleep and awake states. (B) Normalized spectrogram in rat mPFC after hippocampal IED onset. The inset shows the change in averaged mPFC power spectrum from before (green) to after (blue) IED. (C) Average peri-event firing-rate histograms of mPFC pyramidal cells in the time window around hippocampal IED (which occurred at time zero) reveals a decrease in activity (‘down time’) in mPFC firing after the IED. Human studies: (D) LFP recorded from IED electrode in parahippocampal gyrus (upper) and subdural cortical electrode (lower) demonstrating time-locked spindle. The grid in frontal cortex shows z-scored spindle-band power across cortical electrocorticography (ECoG) array triggered on IED (white channels are non-functional). (E) ECoG grid and strip placement (each square represents one recording electrode) on the projected pial surface of four patients with epilepsy. Warm colours indicate high IED-spindle correlation; cool colours represent low correlation. Modified with permission from Gelinas et al. (2016).

Figure 5

Default mode network and epilepsy. (A) DMN regions show decreases in activity when subjects perform cognitive tasks performance. (B) BOLD resting state activity is strongly correlated within DMN regions. Here activity is shown for the seed region in posterior cingulate cortex (yellow arrow in A) and another region which shows a similar pattern of activity, in medial prefrontal cortex (orange arrow in A). (C) Functional connectivity across DMN regions defined by spatial coherence in resting state BOLD signal fluctuations across these areas. (D) Top panel shows scalp EEG with right temporal IED and associated BOLD changes in a patient with TLE. Functional MRI demonstrates significant simultaneous activations in right frontal and parietal lobes and deactivation in cuneus. Bottom panel shows stereotaxic intracerebral EEG (SEEG) traces illustrating runs of hippocampal IEDs (lower three traces) and SEEG in posterior cingulate cortex (PCC), inferior parietal lobule (IPL) and anterior cingulate cortex (ACC). The most medial channel in PCC (upper trace) shows propagation of epileptic activity. (A–C) Adapted with permission from Raichle (2015); (D) Adapted from Fahoum et al. (2013).