Browsing by Author "Nurmi, Tarmo"
Now showing 1 - 6 of 6
- Results Per Page
- Sort Options
- Construction and multilayer motif analysis of temporal fMRI brain networks
Perustieteiden korkeakoulu | Master's thesis(2019-08-19) Nurmi, TarmoMultilayer networks are a tool for incorporating temporal information or different types of data into a single mathematical network object, and with the pipeline developed in this thesis, can be analysed for statistically enriched equivalence classes of isomorphic subnetworks, called motifs, in order to understand the structure of the network. The function of the human brain can be represented as a network of brain regions, Regions of Interest (ROIs). In this thesis, I describe and implement a pipeline for constructing multilayer temporal brain networks (layers correspond to time windows) and analysing their motifs from fMRI (functional magnetic resonance imaging) data. For a given fMRI time series data, ROIs are usually defined as unchanging in time, which, according to recent literature, might be an inaccurate assertion. In the present pipeline, ROIs within each layer can be defined independently, resulting in ROIs that change in time. The pipeline finds isomorphism class distributions of the connected, induced subnetworks of the multilayer brain networks and compares them between groups of data to identify motifs. As a proof-of-concept, the pipeline is applied to a real-world fMRI data set, collected at Aalto University Department of Neuroscience and Biomedical Engineering, which consists of subjects from Finnish and Russian cultural backgrounds listening to a spoken story with culture-specific elements. We find no points in time where the isomorphism class distributions of the two subject groups differ statistically significantly. However, we confirm the expected result that, in general, the isomorphism class counts in human brains are statistically significantly different from those in randomized multilayer Erdős-Rényi networks. The pipeline is therefore shown to find motifs related to fundamental human brain function, but identifying the possibly subtle differences between the two subject groups might require more specific hypotheses (restricting the number of isomorphism classes studied) or more data. - Graphlets in multilayer networks
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä(2022-04-06) Sallmen, Sallamari; Nurmi, Tarmo; Kivelä, MikkoRepresenting various networked data as multiplex networks, networks of networks and other multilayer networks can reveal completely new types of structures in these systems. We introduce a general and principled graphlet framework for multilayer networks which allows one to break any multilayer network into small multilayered building blocks. These multilayer graphlets can be either analysed themselves or used to do tasks such as comparing different systems. The method is flexible in terms of multilayer isomorphism, automorphism orbit definition and the type of multilayer network. We illustrate our method for multiplex networks and show how it can be used to distinguish networks produced with multiple models from each other in an unsupervised way. In addition, we include an automatic way of generating the hundreds of dependency equations between the orbit counts needed to remove redundant orbit counts. The framework introduced here allows one to analyse multilayer networks with versatile semantics, and these methods can thus be used to analyse the structural building blocks of myriad multilayer networks. - Mikrofluidistiset elinmallit lääketestauksen työkaluna
Sähkötekniikan korkeakoulu | Bachelor's thesis(2016-05-04) Nurmi, Tarmo - Molecular Resolution of the Water Interface at an Alkali Halide with Terraces and Steps
A2 Katsausartikkeli tieteellisessä aikakauslehdessä(2016-09-08) Ito, Fumiaki; Kobayashi, Kei; Spijker, Peter; Zivanovic, Lidija; Umeda, Kenichi; Nurmi, Tarmo; Holmberg, Nico; Laasonen, Kari; Foster, Adam S.; Yamada, HirofumiHydration structures at crystal surfaces play important roles in crystal growth or dissolution processes in liquid environments. Recently developed two-dimensional (2D) and three-dimensional (3D) force mapping techniques using frequency-modulation atomic force microscopy (FM-AFM) allow us to visualize the hydration structures at the solid-liquid interfaces at angstrom-scale resolution in real space. Up to now, the experimental and theoretical studies on local hydration structures have mainly focused on those on the terrace, but little work has looked at step edges, usually the key areas in dissolution and growth. In this study, we measured local hydration structures on water-soluble alkali halide crystal surfaces by 2D force mapping FM-AFM. The atomic-scale hydration structures observed on the terraces agree well with molecular-dynamics (MD) simulations. We also measured the hydration structures at the step edge of the NaCl(001) surface, which was constantly dissolving and growing, leading to the clear observation of atomic fluctuations. We found, with the support of MD simulations, that the hydration structures measured by FM-AFM at a time scale of a minute can be interpreted as the time-average of the hydration structures on the upper terrace and those on the lower terrace. - pymnet: A Python Library for Multilayer Networks
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä(2024) Nurmi, Tarmo; Badie Modiri, Arash; Coupette, Corinna; Kivelä, MikkoMany complex systems can be readily modeled as networks and represented as graphs. Such systems include social interactions, transport infrastructures, biological pathways, brains, ecosystems, and many more. A major advantage of representing complex systems as graphs is that the same graph tools and methods can be applied in a wide variety of domains. However, the graph representation has its limitations: many systems contain nodes with multidimensional features, interactions of various types, different levels of hierarchy, or multiple modalities, which deserve to be modeled but cannot be described by simple graphs. Multilayer networks (Kivelä et al., 2014) generalize graphs to capture the rich network data often associated with complex systems, allowing us to study a broad range of phenomena using the same representations, tools, and methods. With pymnet, we introduce a Python package that provides the essential data structures and computational tools for multilayer-network analysis. As highlights, the library offers efficient and scalable implementations for sparse multilayer networks and multiplex networks, integration with bliss to analyze multilayernetwork isomorphisms and automorphisms, and versatile methods for multilayer-network visualization. - Subnetwork enumeration algorithms for multilayer networks
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä(2024) Nurmi, Tarmo; Kivela, MikkoTo understand the structure of a network, it can be useful to break it down into its constituent pieces. This is the approach taken in a multitude of successful network analysis methods, such as motif analysis. These methods require one to enumerate or sample small connected subgraphs of a network. Efficient algorithms exists for both enumeration and uniform sampling of subgraphs, and here we generalize the ESU algorithm for a very general notion of multilayer networks. We show that multilayer network subnetwork enumeration introduces nontrivial complications to the existing algorithm, and present two different generalized algorithms that preserve the desired features of unbiased sampling and scalable, communicationfree parallelization. In addition, we introduce a straightforward aggregation-disaggregation-based enumeration algorithm that leverages existing subgraph enumeration algorithms. We evaluate these algorithms in synthetic networks and with real-world data, and show that none of the algorithms is strictly more efficient but rather the choice depends on the features of the data.