352
K. Lehnertz
the various temporal and spatial scales assessed with the different computational
models in order to finely balance model simplification with biological realism.
Epilepsy is considered a dynamical brain disease [103], and the transition to
seizure is often conceptualized as a (critical) phase transition [68, 104]. Such a
transition can be heralded by the phenomenon of critical slowing down, and the
human brain is among the many natural systems in which critical slowing down has
been repeatedly claimed to provide early warning signals for transitions into (and
out of) epileptic seizures (see, e.g., [100] and references therein). Recent studies
[102, 164], however, could not corroborate these claims when evaluating the predictive performance (cf. Sect. 22.3) of widely used early warning indicators (variance,
lag-1 autocorrelation) on large samples. Thus the assumed mechanism behind the
critical transition (bifurcation-induced tipping) may be too simplistic for the human
epileptic brain which calls for techniques that allow identification of early warning indicators for other transition scenarios [12, 133]. Switching between different
states may emerge as a result of multistability of the brain [65] and recent findings
indicate that switching might also be induced by changes in the gross connections
of the neuronal network [9] and not only by altered properties of neurons or groups
of neurons. Notwithstanding the high relevance of improving our understanding of
mechanisms underlying the transition to seizure, future studies would also need to
address the mechanism underlying the transition to the preseizure state.
The probably most important—but as yet unsolved issue—of prediction and prevention of epileptic seizures centers around the controllability of the human epileptic
brain. It is a nonlinear, open, dissipative and adaptive system, innately designed to
learn. Learning is not only tightly related to neuronal plasticity [53] but also linked to
brain disorders such as epilepsy [18]. One might thus speculate that seizures present
an abnormal learned response to recurrent perturbations—such as seizures [54]. If
epilepsy is indeed a “earned” disease, it will be a challenging endeavor to identify
powerful control strategies to prevent the epileptic brain from generating seizures.
Part of this difficulty may be attributed to the brain’s stability properties with respect
to the aforementioned endogenous and/or exogenous inputs. Indeed, recent findings indicate that in many subjects with epilepsy brain resilience increases rather
than decreases prior to seizures and that this preseizure increase clearly exceeds
physiologically induced fluctuations of brain resilience [131]. Research along these
lines is urgently needed to better understand how, when, and why the epileptic brain
efficiently defies control by virtue of its intrinsic plasticity and adaptiveness.
References
1. A. Aarabi, B. He, Seizure prediction in patients with focal hippocampal epilepsy. Clin. Neurophysiol. 128, 1299–1307 (2017)
2. U.R. Acharya, Y. Hagiwara, H. Adeli, Automated seizure prediction. Epilepsy Behav. 88,
251–261 (2018)
3. R. Albert, A.L. Barabási, Statistical mechanics of complex networks. Rev. Mod. Phys. 74,
47–97 (2002). https://doi.org/10.1103/RevModPhys.74.47
K. Lehnertz
the various temporal and spatial scales assessed with the different computational
models in order to finely balance model simplification with biological realism.
Epilepsy is considered a dynamical brain disease [103], and the transition to
seizure is often conceptualized as a (critical) phase transition [68, 104]. Such a
transition can be heralded by the phenomenon of critical slowing down, and the
human brain is among the many natural systems in which critical slowing down has
been repeatedly claimed to provide early warning signals for transitions into (and
out of) epileptic seizures (see, e.g., [100] and references therein). Recent studies
[102, 164], however, could not corroborate these claims when evaluating the predictive performance (cf. Sect. 22.3) of widely used early warning indicators (variance,
lag-1 autocorrelation) on large samples. Thus the assumed mechanism behind the
critical transition (bifurcation-induced tipping) may be too simplistic for the human
epileptic brain which calls for techniques that allow identification of early warning indicators for other transition scenarios [12, 133]. Switching between different
states may emerge as a result of multistability of the brain [65] and recent findings
indicate that switching might also be induced by changes in the gross connections
of the neuronal network [9] and not only by altered properties of neurons or groups
of neurons. Notwithstanding the high relevance of improving our understanding of
mechanisms underlying the transition to seizure, future studies would also need to
address the mechanism underlying the transition to the preseizure state.
The probably most important—but as yet unsolved issue—of prediction and prevention of epileptic seizures centers around the controllability of the human epileptic
brain. It is a nonlinear, open, dissipative and adaptive system, innately designed to
learn. Learning is not only tightly related to neuronal plasticity [53] but also linked to
brain disorders such as epilepsy [18]. One might thus speculate that seizures present
an abnormal learned response to recurrent perturbations—such as seizures [54]. If
epilepsy is indeed a “earned” disease, it will be a challenging endeavor to identify
powerful control strategies to prevent the epileptic brain from generating seizures.
Part of this difficulty may be attributed to the brain’s stability properties with respect
to the aforementioned endogenous and/or exogenous inputs. Indeed, recent findings indicate that in many subjects with epilepsy brain resilience increases rather
than decreases prior to seizures and that this preseizure increase clearly exceeds
physiologically induced fluctuations of brain resilience [131]. Research along these
lines is urgently needed to better understand how, when, and why the epileptic brain
efficiently defies control by virtue of its intrinsic plasticity and adaptiveness.
References
1. A. Aarabi, B. He, Seizure prediction in patients with focal hippocampal epilepsy. Clin. Neurophysiol. 128, 1299–1307 (2017)
2. U.R. Acharya, Y. Hagiwara, H. Adeli, Automated seizure prediction. Epilepsy Behav. 88,
251–261 (2018)
3. R. Albert, A.L. Barabási, Statistical mechanics of complex networks. Rev. Mod. Phys. 74,
47–97 (2002). https://doi.org/10.1103/RevModPhys.74.47