Pulse desynchronization of neural populations by targeting the centroid of the limit cycle in phase space
Ramón Guevara, Marco Zenari, Giorgio Nicoletti, Elisa Marini, Samir Suweis, Sandro Azaele, Marco Formentin · 2026
Read Pulse desynchronization of neural populations by targeting the centroid of the limit cycle in phase space by Ramón Guevara, Marco Zenari, Giorgio Nicoletti, Elisa Marini, Samir Suweis, Sandro Azaele, Marco Formentin (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-12878.txt.
Introduction
Ramón Guevara and co-authors propose a way to break up harmful synchrony in neural populations with brief pulses. Over-synchronized activity marks conditions like epilepsy and Parkinson disease. When the true system is unknown, picking pulse times and strengths is a hard inverse problem. Their scheme targets the centroid of the limit cycle in phase space. That point is easy to compute and stays robust as coupling changes. It also sits where return times are shortest, which aids control.
Catalog record
| CREATOR | Ramón Guevara, Marco Zenari, Giorgio Nicoletti, Elisa Marini, Samir Suweis, Sandro Azaele, Marco Formentin |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35421 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.12878
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Pulse desynchronization of neural populations by targeting the centroid of the limit cycle in phase space Ramón Guevara, Marco Zenari, Giorgio Nicoletti, Elisa Marini, Samir Suweis, Sandro Azaele, Marco Formentin Abstract The synchronized activity of neuronal populations can lead to pathological over-synchronization in conditions such as epilepsy and Parkinson disease. Such states can be desynchronized by brief electrical pulses. But when the underlying oscillating system is not known, as in most practical applications, to determine the specific times and intensities of pulses used for desynchronizaton is a difficult inverse problem. Here we propose a desynchronization scheme for neuronal models of bi-variate neural activity, with possible applications in the medical setting. Our main argument is the existence of a peculiar point in the phase space of the system, the centroid, that is both easy to calculate and robust under changes in the coupling constant. This important target point can be used in a control procedure because it lies in the region of minimal return times of the system. 1 Introduction The adjustment of rhythms of self-sustained or weakly chaotic oscillators due to weak coupling is a widespread phenomenon in nature, with examples ranging from physics to chemistry, biology (e.g., pacemaker cardiac cells, Josephson junctions, lasers, mechanical vibrations, electronic circuits, metronomes, fireflies) ( 30 ). Understanding network synchronization is particularly relevant for neuroscience ( 9 ) because the collective coordinated (synchronized) activity of neuronal populations is at the origin of neural rhythms ( 4 ). Furthermore, neural synchronization is believed to play a fundamental role in brain functionality, as a mechanism for information integration in perceptual and cognitive processes ( 4 ; 35 ; 43 ). The malfunction of synchronization mechanisms may also affect brain functionality. It is well known that excessive neural synchronization is often associated with pathological activity, particularly in epilepsy, Parkinson’s disease, and essential tremor ( 13 ; 29 ; 1 ; 25 ). For example, during an epileptic seizure, the activity of neuronal networks is highly synchronized across brain areas, leading to a well-recognized high amplitude and fairly rhythmic electroencephalographic (EEG) signal. For this reason, the last few decades have witnessed increased interest in the monitoring and control of neuronal synchronization, with the ultimate goal of suppressing pathological over-synchronized brain states. Recent technological advances have provided a way to alleviate such pathological conditions by low-intensity electric stimulation of the affected brain regions or peripheral nerves. This is particularly useful for patients who do not respond to drug medications. In this sense, one of the most important breakthroughs was the development of deep brain stimulation (DBS), a technique to suppress synchronized activity in epilepsy and Parkinson disease ( 2 ; 15 ; 17 ). In DBS, micro-electrodes implanted in the brain locally stimulate neuronal tissue by application of high frequency electric pulses (around 120 Hz), effectively modulating neuronal dynamics. The mechanism of action of DBS is not yet fully understood ( 14 ; 7 ). However, it has been hypothesized that it suppresses pathological oscillatory activity through the desynchronization of neuronal populations, without suppressing the oscillatory activity of individual neurons ( 39 ). Due to the negative effects that injected currents can induce in patients (such as speech impairment and involuntary muscle contraction) ( 28 ; 36 ), the current goal of DBS and similar techniques is to achieve a minimally invasive stimulating signal, that is, to achieve desynchronization of neuronal populations with a minimal number of injected pulses. Two fundamental approaches to achieve desynchronization of collective neural oscillations have been proposed in the literature: 1) those based on phase resetting techniques ( 37 ; 38 ; 39 ; 11 ; 20 ) and the injection of small pulses of current; 2) Those originating from the control theory community, using continuous close-loop feedback stimulation ( 31 ; 32 ; 26 ). The latter approach, based on the idea of using a closed loop feedback control of synchronization, has acquired great credibility due to its effectiveness in theoretical models and simulations. In fact, it offers the possibility of implementing a vanishingly small stimulation signal, a crucial feature in medical applications (as we mentioned above, DBS can lead to collateral damages, so it is important to minimize the stimulating signal) ( 42 ; 26 ). In the medical setting, the closed loop feedback stimulation concept is also gaining credibility and has been implemented with relative success ( 34 ; 19 ). In more recent years, the idea of a desynchronization of neuronal populations by closed-loop feedback using stimulating pulses instead of continuous signals (that is, a combination of both approaches) was developed and applied to neuronal models, showing that it is possible to desynchronize the output activity of the models by very small-amplitude pulses ( 33 ; 16 ). In the current study, we follow a similar approach, that of synchronization control through pulse stimulation, but with two main differences from previous work ( 33 ; 16 ). The first difference has to do with the input signal (input to the feedback-control system). That is, what kind of neurophysiological signal does our device measure? In previous literature, the input is univariate, but we consider the case where a bi-variate signal is measured. Indeed, it is typically assumed that the input signal is the local field potential (average voltage of the neuronal population, a univariate time series) or a related signal, as is customary in electrophysiology. Instead, here we assume that a bi-variate time series is measured, which is justified by recent advances in neurophsyological and multimodal measurements, allowing for bi-variate, and more generally multivariate recordings. For example, at the level of local circuits, it is possible to simultaneously record neuronal spikes and calcium imaging in vivo ( 44 ), an example that would constitute a bi-variate input signal for a control device. Recent works have also shown that it is possible to measure excitation and inhibition (that is, a bi-variate signal) simultaneously, on the basis of electrophysiological recordings ( 27 ; 3 ). Another example of a multivariate signal is the simultaneous measurement of brain activity using EEG and MEG (although it is worth remembering that EEG and magnetoencephalography (MEG) signals are only partially independent) ( 22 ). We highlight that, at the whole brain scale, especially when such non-invasive techniques (scalp EEG or MEG) are used, oscillatory activity is described in terms of phenomenological models that do not make explicit use of individual neurons ( 10 ). In this direction, Wilson-Cowan population models have been extensively used to model medium and large-scale brain activity in terms of excitation and inhibition ( 8 ; 5 ). Such models have been successfully used to describe epileptic seizures ( 24 ; 21 ). Second, the control problem is addressed in previous work through successive approximations; that is, by an adaptive search for the most effective phase for pulse stimulation ( 33 ). This strategy requires applying many pulses before the population signal is sufficiently small. Here, we exploit the fact that the input signal is bi-variate to extract information about the phase space and other dynamic and geometric properties of the underlying dynamical system, without completely fitting it to the data. Knowledge of such properties gives us the opportunity to achieve synchronization control