20 November 2012. A new study published in Neuron on November 8 finds that sleep architecture is altered in a rat neurodevelopmental model proposed to mimic some of the features of schizophrenia. A team of researchers led by Matt Jones of Eli Lilly and Company in Surrey, United Kingdom, describes a coordinated relationship between different types of cortical and hippocampal oscillations in non-REM (NREM) sleep in control rats, but demonstrates a decoupling of these oscillations in rats treated in utero with methylazoxymethanol-acetate (MAM).
Sleep disturbances are common in schizophrenia, and include alterations in both rapid eye movement (REM) and NREM sleep (Wulff et al., 2010). In fact, improved quality of sleep is rated as an important treatment goal by patients (Auslander and Jeste, 2002). Abnormal sleep architecture is correlated with cognitive deficits in schizophrenia, leading to the hypothesis that sleep problems may, in part, underlie the impaired cognition (Manoach and Stickgold, 2009).
Individual oscillation properties are preserved
In the current study, first authors Keith Phillips and Ullrich Bartsch investigated several prominent oscillations of NREM sleep that underlie memory consolidation—cortical delta waves (0.3-3 Hz), thalamocortical sleep spindles (10-15 Hz), and hippocampal ripples (120-250 Hz) (Diekelmann and Born, 2010)—using electroencephalography, local field potential, and unit recordings of rats. They studied the sleep architecture of two groups: control rats and those exposed to the mitotoxin MAM during embryonic development, a proposed neurodevelopmental model for schizophrenia. MAM treatment interferes with the division and migration of neurons, and researchers can target different cortical circuits by varying the dose and timing of delivery. Rats treated with MAM on embryonic day 17 show several features reminiscent of schizophrenia, including cognitive impairments, as well as glutamatergic and GABAergic dysfunction (Lodge and Grace, 2009; see also SRF related news story; SRF news story).
In the current study, MAM-17 rats displayed normal circadian rhythms and REM sleep, but showed a reduction in their amount of NREM sleep. The researchers next examined the neurophysiological features of this impaired NREM sleep, finding reduced delta wave power, and spindle density and amplitude in MAM rats at visual cortex recording sites. In contrast to these posterior abnormalities, the mechanisms of delta wave and spindle generation in motor cortex, as well as hippocampal ripple properties, were largely unaffected by MAM treatment.
Relationship between oscillations is altered
The researchers next analyzed the temporal relationship between oscillations. In control rats, delta wave coherence was observed between the motor and visual cortical electrodes, with the motor cortical delta waves preceding the visual cortical ones, consistent with the anteroposterior direction of travel reported previously (Massimini et al., 2004). However, this coherence was reduced in MAM-treated rats, suggesting an impaired propagation and synchrony of delta waves.
The phase locking, or timing, of thalamocortical spindles relative to delta waves was also impaired in MAM rats at motor cortex electrode sites, but preserved at visual cortex sites, consistent with a posterior disruption of network activity. In addition, the coordination between hippocampal ripples and prelimbic cortical spindles was disrupted in the MAM group. In the control group, ripple power peaks preceded those of spindles, and spindle power modulated ripple power, consistent with a prior study (Clemens et al., 2011). However, this modulation was either completely lacking or greatly reduced in MAM rats.
Phillips, Bartsch and colleagues then examined whether the spike timing of single units was altered by MAM treatment, finding that the timing of prelimbic cortical and hippocampal units was shifted, and that prelimbic cortical units were less phase locked to spindles in MAM rats than controls. In addition, in control rats, those units with the strongest spindle phase locking fired more spikes during ripples than the units that were not as strongly phase locked. However, this relationship was absent in MAM-treated rats, further suggesting reduced ripple-spindle coordination and hippocampal-prelimbic cortical decoupling during NREM sleep in MAM rats.
Thus, MAM rats exhibit impaired propagation of cortical delta waves and incorrectly timed spindle oscillations during NREM sleep, resulting in impaired coordination with hippocampal ripples and altered spike timing. As for a mechanism underlying the decoupling of oscillations during NREM sleep, the authors suggest that the abnormalities in parvalbumin interneurons observed both in schizophrenia and the MAM model may play a role, since these cells are involved in the generation of cortical oscillations (Lodge et al., 2009; Gonzalez-Burgos et al., 2011). The authors conclude that “our study serves to emphasize that disrupted thalamic-cortical-limbic network activity during sleep must therefore be considered alongside waking activity as a therapeutic target in schizophrenia and related diseases.”—Allison A. Curley.
Phillips KG, Bartsch U, McCarthy AP, Edgar DM, Tricklebank MD, Wafford KA, Jones MW. Decoupling of sleep-dependent cortical and hippocampal interactions in a neurodevelopmental model of schizophrenia. Neuron. 2012 Nov 8 ;76(3):526-33. Abstract