% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@PHDTHESIS{Zajzon:1010731,
author = {Zajzon, Barna},
othercontributors = {Morrison, Abigail Joanna Rhodes and Schaub, Michael Thomas},
title = {{S}equential information processing in modular spiking
networks},
volume = {56},
school = {RWTH Aachen University},
type = {Dissertation},
address = {Düren},
publisher = {Shaker Verlag},
reportid = {RWTH-2025-04286},
isbn = {3-8440-9957-3},
series = {Aachener Informatik Berichte, Software Engineering},
pages = {1 Online-Ressource : Illustrationen},
year = {2025},
note = {Druckausgabe: 2025. - Auch veröffentlicht auf dem
Publikationsserver der RWTH Aachen University; Dissertation,
RWTH Aachen University, 2024},
abstract = {Molded by evolutionary processes to cope with the
statistical regularities in the world, the symbiotic
relation between the structure, dynamics and function of the
neural machinery underlies all behavioral and cognitive
processes. Established paradigms formulate these processes
in terms that involve the manipulation of sequentially
organized time-discrete (symbolic) representations. This
underscores two basic functional requirements that cortical
circuits must fulfill: the ability to create suitable
representations from a highly volatile and noisy
environment; and the capacity to process, and learn from,
their spatio-temporal structure. While the precise
mechanisms are largely unknown, these processes must be
implemented in the biophysical substrate of the brain, where
the complex interactions of neuronal populations can
leverage a hierarchical and modular architecture in order to
process information on multiple spatial and temporal scales.
From a modeler’s perspective, one can tackle these
problems from two complementary angles: identify some
fundamental organizing principles, such as modularity, and
try to elucidate their role (bottom-up); or focus on a
specific functionality, like sequence processing, and devise
possible, biophysically plausible models for it (top-down).
Combining software tools, simulation studies and theoretical
analysis, this thesis touches upon both approaches over the
course of a series of research projects, with the shared
goal of disentangling how modular structures enable neural
circuits to learn and process sequential information in an
efficient and reliable manner. The first part analyses the
characteristics of state representations in modular spiking
networks and the architectural and dynamical constraints
that influence the system’s ability to retain, transfer
and integrate stimulus information in the presence of noise.
It explores the novel hypothesis that modular topographic
maps, a pervasive anatomical feature of the cortex, may
provide a structural scaffold for sequential denoising of
stimulus representations. By combining modeling with network
theory, this thesis demonstrates that the sharpness of
topographic projections acts as a bifurcation parameter,
controlling the macroscopic dynamics and representational
precision of the system. In-depth theoretical analysis
unravels the dynamical principles underlying the mechanism,
and suggests a robust and generic structural feature that
enables a broad range of behaviorally-relevant operating
regimes. The second part of this work is dedicated to
investigating existing, biologically detailed models of
sequence processing. If we are to harvest the knowledge
within these models and arrive at a deeper mechanistic
understanding of the involved phenomena, it is critical that
the models and their findings are accessible, reproducible,
and quantitatively comparable. First, the importance of
these aspects are illustrated through a replication study.
Building on this, the study lays the initial steps towards a
conceptual and practical, theoretically-grounded framework
for benchmarking and comparison of sequence learning models.
Through such a meta-analysis study, it aims not only to
provide critical evaluation of current models, but also to
synthesize their insights into a set of functional and
neurobiological features that could be corroborated with
experimental data and guide future studies.},
cin = {124920 / 120000 / 535000-7 ; 935810 / 121510},
ddc = {004},
cid = {$I:(DE-82)124920_20200227$ / $I:(DE-82)120000_20140620$ /
$I:(DE-82)535000-7_20140620$ / $I:(DE-82)121510_20140620$},
pnm = {neuroIC002 - Recurrence and stochasticity for
neuro-inspired computation (EXS-SF-neuroIC002) / SMHB -
Supercomputing and Modelling for the Human Brain
(HGF-SMHB-2013-2017) / DFG project G:(GEPRIS)491111487 -
Open-Access-Publikationskosten / 2025 - 2027 /
Forschungszentrum Jülich (OAPKFZJ) (491111487) /
Exploratory Research Space: Seed Fund (2) als
Anschubfinanzierung zur Erforschung neuer
interdisziplinärer Ideen (EXS-SF)},
pid = {G:(DE-82)EXS-SF-neuroIC002 / G:(DE-Juel1)HGF-SMHB-2013-2017
/ G:(GEPRIS)491111487 / G:(DE-82)EXS-SF},
typ = {PUB:(DE-HGF)11 / PUB:(DE-HGF)3},
doi = {10.18154/RWTH-2025-04286},
url = {https://publications.rwth-aachen.de/record/1010731},
}
guest ::