Authors :
Presenting Author: Alex Bender, MD, PhD – Massachusetts General Hospital, Harvard Medical School
William Leung, MBBS – Massachusetts General Hospital
Kyle Pellerin, BS – Massachusetts General Hospital
Maddie Taubkin, MPH – Massachusetts General Hospital
Alice Lam, MD, PhD – Harvard
Rationale:
Cognitive impairment is a common and debilitating comorbidity in epilepsy. Spindles, along with slow oscillations (SO), are distinct graphoelements of the NREM sleep EEG microarchitecture that are theorized to play active roles in memory consolidation, and we recently showed that these are altered in adult temporal lobe epilepsy (ref: Bender et al 2023 Neurology). Understanding how changes in these neurophysiological parameters relate to cognitive function may reveal insights into the mechanisms of cognitive impairment in epilepsy and serve as potential biomarkers for cognitive decline.
Methods:
We prospectively studied adult patients admitted to the Epilepsy Monitoring Unit (EMU) at Massachusetts General Hospital who did not have any prior brain surgery, tumor or known developmental intellectual disability. A global cognitive assessment was completed using the Montreal Cognitive Assessment (MOCA). In addition, The Rey Auditory Verbal Learning Test (RAVLT), a 15-item word list, was administered with a 30-min short-delay recall and an overnight long-term recall (LTR). Continuous scalp EEG recorded during the same overnight period was obtained in all subjects and pre-processed using previously published algorithms for automated sleep staging (ref: Jaoude et al 2020 Sleep), as well as detection of sleep spindles and slow oscillation signals. Fronto-temporal derivations (F3, T5) were chosen for analysis based on results of a prior study revealing deficits in these regions in epilepsy. Associations of sleep parameters with cognitive performance were then tested with linear regression analyses.
Results:
We enrolled 43 subjects of mean age 39 years (range 18-71). Scores on the MOCA ranged from 17-30 with a mean of 25 (std dev 3.1). On the RAVLT, subjects recalled a mean of 9 words (std dev 4.1) at a 30-minute delay and 8 words (std dev 1.4) after an overnight delay. Spindle density in frontal (F3) and temporal (T5) regions was positively associated with MOCA (F3: R=0.40, p=0.019; T5: R=0.44, p=0.009) and overnight LTR (F3: R=0.35, p=0.048; T5: R=0.47, p=0.006). The mean phase of spindle-SO coupling in frontal regions was also associated with both MOCA (F3: R=0.39, p=0.022) and overnight LTR (F3: R=0.45, p=0.009). However, in temporal regions, the phase value was only associated with overnight LTR (T5: R=-0.38, p=0.029), rather than MOCA (T5: R=0.015, p=0.93), and demonstrated an inverse phase relationship in comparison to coupling in frontal regions. Other sleep microarchitectural parameters, including the spindle-SO coupling strength and SO power were not associated with cognitive performance on either test.
Conclusions:
In a prospective cohort of patients admitted to the EMU, fronto-temporal spindle density and spindle-SO coupling phase were associated with both global cognition and overnight LTR. We postulate that spindle density and spindle-SO coupling phase are not only important for memory consolidation processes, but more generally reflect on the health of brain networks in sleep. In this regard, they may serve as useful neurophysiological biomarkers for cognition in epilepsy.
Funding:
NINDS R25NS065743