Introduction
Higher order brain functions rely on distributed networks involving the
cortex and subcortical regions. The thalamus plays a crucial role in the
proper functioning of these networks: thalamo-cortical loops are
associated with sensory processing, memory formation, executive
functions, many of which are tightly related with sleep state-dependent
neural oscillations (Fama and Sullivan 2015; Wolff et al. 2021). The
thalamus also plays an important role in the coordinated connection
between the cortex and hippocampus (Latchoumane et al. 2017). The human
anterior thalamic nucleus (ANT) is a higher-order thalamic nucleus which
is interconnected with the hippocampus (Aggleton et al. 2010). It has
reciprocal connections with the anterior cingulate cortex, retrosplenial
cortex, and subiculum, and selective inactivation of the ANT leads to
impaired memory formation similarly to hippocampal lesions. Furthermore,
the ANT plays an important role in the propagation of epileptic seizures
and therefore became an important target for deep brain stimulation
(DBS) which serves as a treatment for medically refractory epilepsy
(Salanova 2018). Another important higher-order nucleus in the human
thalamus is the mediodorsal nucleus (MD) which is reciprocally connected
with the medial prefrontal cortex (mPFC) and also receives inputs from
parahippocampal regions, and therefore it is assumed to interact with
the cortex and hippocampus in declarative memory formation during
non-REM (NREM) sleep (Mitchell and Chakraborty 2013). Reduced MD volume
was associated with decreased sleep spindle density in frontal brain
regions which suggest that the MD is involved in sleep spindle
generation and/or propagation (Buchmann et al. 2014).
Sleep spindles are NREM sleep state-specific oscillations characterized
by waxing/waning 10–16 Hz waveforms that are generated by the reticular
nucleus of the thalamus and are propagated to other brain regions by
thalamo-cortical circuits (Huguenard and McCormick 2007; Steriade 2005).
Animal studies confirmed that dynamic reticular
thalamic-thalamo-cortical interactions are responsible for the
generation of sleep spindles (Steriade, 2005; Steriade et al., 1987),
which were associated with cognitive functions such as learning and
memory consolidation (Cairney et al. 2018; Saletin, Goldstein, and
Walker 2011), and intellectual ability (Fogel and Smith 2011; Ujma 2018;
Ujma, Bódizs, and Dresler 2020). Sleep spindles tend to co-occur with
and coordinate the faster oscillations, called hippocampal
ripples (~100–200 Hz), providing efficient off-line
plasticity windows for memory consolidation and reorganization
(Girardeau and Zugaro 2011; Wilson and McNaughton 1994). Besides neural
plasticity sleep spindles were associated with pathological off-line
plasticity in epileptic models and human epilepsy patients (Gelinas et
al. 2016). Findings regarding the strong relationship between NREM sleep
oscillations, neuroplasticity and epilepsy indicate that
sleep-associated interictal epileptic discharges (IEDs) harm cognitive
functions, inducing a significant cognitive loss (Halász et al. 2019).
The aim of the current study was to assess the function of ANT and MD in
sleep-related neural oscillations measured by the association between
thalamic sleep spindles and thalamic ripples during NREM sleep. It was
proposed that sleep-related epileptic transformation of normal
neurological networks, involving the hippocampus, thalamus and cortex,
interfere with sleep-related synaptic homeostasis, neural plasticity,
and cognitive functioning (for a recent review, see Halász and Szűcs
2020). In this pathway, the role of the human ANT and MD is not clear.
Although, different aspects of sleep spindle-related dynamic
thalamocortical interactions were revealed (Mak-Mccully et al. 2017;
Tsai et al. 2010), relationships between sleep spindles and ripples in
the thalamus remained questionable. A recent report (Rektor et al. 2016)
indicated the occurrence of high frequency oscillations, such as ripples
in the human thalamus, however results confirming the involvement of
ripples in the thalamus is remarkably sparse. Furthermore, IEDs in the
human thalamus, such as in the ANT LFP records were observed and
suggested the involvement of ANT in the propagation of epileptic
activity and to contribute to the epileptic network (Hodaie et al. 2002;
Sweeney-Reed et al. 2016). Here we investigated the association between
thalamic sleep spindles and ripples within two higher-order thalamic
nuclei in human subjects, the ANT and the MD. We hypothesized
that sleep spindles and ripples in these thalamic nuclei are not
pathological expressions of epilepsy, but physiological oscillatory
patterns. Sleep spindles, and especially fast sleep spindles measured on
the scalp have a close connection with cognitive performance and
intelligence (Bestmann et al. 2019; Bódizs et al. 2005; Chatburn et al.
2013; Ujma 2018). The ANT has tight anatomical and functional
connections with the hippocampus, as well as different parts of the
prefrontal cortex, and might act as an interface between them. In order
to test the hypothesis that sleep spindles and ripples are physiological
oscillations in the ANT, we assessed the association between sleep
spindles and ripples with general intelligence and clinical epilepsy
characteristics.