It is a widely accepted fact that the computational capability of recurrent neural networks (RNNs) is maximized on the so-called "edge of criticality." Once the network operates in this ...configuration, it performs efficiently on a specific application both in terms of: 1) low prediction error and 2) high short-term memory capacity. Since the behavior of recurrent networks is strongly influenced by the particular input signal driving the dynamics, a universal, application-independent method for determining the edge of criticality is still missing. In this paper, we aim at addressing this issue by proposing a theoretically motivated, unsupervised method based on Fisher information for determining the edge of criticality in RNNs. It is proved that Fisher information is maximized for (finite-size) systems operating in such critical regions. However, Fisher information is notoriously difficult to compute and requires the analytic form of the probability density function ruling the system behavior. This paper takes advantage of a recently developed nonparametric estimator of the Fisher information matrix and provides a method to determine the critical region of echo state networks (ESNs), a particular class of recurrent networks. The considered control parameters, which indirectly affect the ESN performance, are explored to identify those configurations lying on the edge of criticality and, as such, maximizing Fisher information and computational performance. Experimental results on benchmarks and real-world data demonstrate the effectiveness of the proposed method.
Graph representations offer powerful and intuitive ways to describe data in a multitude of application domains. Here, we consider stochastic processes generating graphs and propose a methodology for ...detecting changes in stationarity of such processes. The methodology is general and considers a process generating attributed graphs with a variable number of vertices/edges, without the need to assume a one-to-one correspondence between vertices at different time steps. The methodology acts by embedding every graph of the stream into a vector domain, where a conventional multivariate change detection procedure can be easily applied. We ground the soundness of our proposal by proving several theoretical results. In addition, we provide a specific implementation of the methodology and evaluate its effectiveness on several detection problems involving attributed graphs representing biological molecules and drawings. Experimental results are contrasted with respect to suitable baseline methods, demonstrating the effectiveness of our approach.
Purpose:
To evaluate the effect of plan parameters on volumetric modulated arc therapy (VMAT) dosimetric accuracy, together with the possibility of scoring plan complexity.
Methods:
142 clinical VMAT ...plans initially optimized using a 4° control point (CP) separation were evaluated. All plans were delivered by a 6 MV Linac to a biplanar diode array for patient-specific quality assurance (QA). Local Γ index analysis (3%, 3 mm and 2%, 2 mm) enabled the comparison between delivered and calculated dose. The following parameters were considered for each plan: average leaf travel (LT), modulation complexity score applied to VMAT (MCSv), MU value, and a multiplicative combination of LT and MCSv (LTMCS). Pearson's correlation analysis was performed between Γ passing rates and each parameter. The effects of CP angular separation on VMAT dosimetric accuracy were also analyzed by focusing on plans with high LT values. Forty out of 142 plans with LT above 350 mm were further optimized using a finer angle spacing (3° or 2°) and Γ analysis was performed. The average Γ passing rates obtained at 4° and at 3°/2° sampling were compared. A further correlation analysis between all parameters and the Γ pass-rates was performed on 142 plans, but including the newly optimized 40 plans (CP every 3° or 2°) in place of the old ones (CP every 4°).
Results:
A moderate significant (p < 0.05) correlation between each examined parameter and Γ passing rates was observed for the original 142 plans at 4° CP discretization. A negative correlation was found for LT with Pearson's r absolute values above 0.6, suggesting that a lower dosimetric accuracy may be expected for higher LT values when a 4° CP sampling is used. A positive correlation was observed for MCSv and LTMCS with r values above 0.5. In order to score plan complexity, threshold values of LTMCS were defined. The average Γ passing rates were significantly higher for the plans created using the finer CP spacing (3°/2°) compared to the plans optimized using the standard 4° spacing (Student t-test p < 0.05). The correlation between LT and passing rates was strongly diminished when plans with finer angular separations were considered, yielding Pearson's r absolute values below 0.45.
Conclusions:
At 4° CP sampling, LT, MCSv, and LTMCS were found to be significantly correlated with VMAT dosimetric accuracy, expressed as Γ pass-rates. These parameters were found to be possible candidates for scoring plan complexity using threshold values. A finer CP separation (3°/2°) led to a significant increase in dosimetric accuracy for plans with high leaf travel values, and to a decrease in correlation between LT and Γ passing rates. These results indicated that the influence of LT on VMAT dosimetric accuracy can be controlled by reducing CP separation. CP spacing for all plans requiring large leaf motion should not exceed 3°. The reported data were integrated to optimize our clinical workflow for plan creation, optimization, selection among rival plans, and patient-specific QA of VMAT treatments.
The space of graphs is often characterized by a nontrivial geometry, which complicates learning and inference in practical applications. A common approach is to use embedding techniques to represent ...graphs as points in a conventional Euclidean space, but non-Euclidean spaces have often been shown to be better suited for embedding graphs. Among these, constant-curvature Riemannian manifolds (CCMs) offer embedding spaces suitable for studying the statistical properties of a graph distribution, as they provide ways to easily compute metric geodesic distances. In this paper, we focus on the problem of detecting changes in stationarity in a stream of attributed graphs. To this end, we introduce a novel change detection framework based on neural networks and CCMs, which takes into account the non-Euclidean nature of graphs. Our contribution in this paper is twofold. First, via a novel approach based on adversarial learning, we compute graph embeddings by training an autoencoder to represent graphs on CCMs. Second, we introduce two novel change detection tests operating on CCMs. We perform experiments on synthetic data, as well as two real-world application scenarios: the detection of epileptic seizures using functional connectivity brain networks and the detection of hostility between two subjects, using human skeletal graphs. Results show that the proposed methods are able to detect even small changes in a graph-generating process, consistently outperforming approaches based on Euclidean embeddings.
Representing patterns as labeled graphs is becoming increasingly common in the broad field of computational intelligence. Accordingly, a wide repertoire of pattern recognition tools, such as ...classifiers and knowledge discovery procedures, are nowadays available and tested for various datasets of labeled graphs. However, the design of effective learning procedures operating in the space of labeled graphs is still a challenging problem, especially from the computational complexity viewpoint. In this paper, we present a major improvement of a general-purpose classifier for graphs, which is conceived on an interplay between dissimilarity representation, clustering, information-theoretic techniques, and evolutionary optimization algorithms. The improvement focuses on a specific key subroutine devised to compress the input data. We prove different theorems which are fundamental to the setting of the parameters controlling such a compression operation. We demonstrate the effectiveness of the resulting classifier by benchmarking the developed variants on well-known datasets of labeled graphs, considering as distinct performance indicators the classification accuracy, computing time, and parsimony in terms of structural complexity of the synthesized classification models. The results show state-of-the-art standards in terms of test set accuracy and a considerable speed-up for what concerns the computing time.
A recurrent neural network (RNN) is a universal approximator of dynamical systems, whose performance often depends on sensitive hyperparameters. Tuning them properly may be difficult and, typically, ...based on a trial-and-error approach. In this work, we adopt a graph-based framework to interpret and characterize internal dynamics of a class of RNNs called echo state networks (ESNs). We design principled unsupervised methods to derive hyperparameters configurations yielding maximal ESN performance, expressed in terms of prediction error and memory capacity. In particular, we propose to model time series generated by each neuron activations with a horizontal visibility graph, whose topological properties have been shown to be related to the underlying system dynamics. Successively, horizontal visibility graphs associated with all neurons become layers of a larger structure called a multiplex. We show that topological properties of such a multiplex reflect important features of ESN dynamics that can be used to guide the tuning of its hyperparamers. Results obtained on several benchmarks and a real-world dataset of telephone call data records show the effectiveness of the proposed methods.
Recurrence quantification analysis of dynamic brain networks Lopes, Marinho A.; Zhang, Jiaxiang; Krzemiński, Dominik ...
European journal of neuroscience/EJN. European journal of neuroscience,
February 2021, Volume:
53, Issue:
4
Journal Article
Peer reviewed
Open access
Evidence suggests that brain network dynamics are a key determinant of brain function and dysfunction. Here we propose a new framework to assess the dynamics of brain networks based on recurrence ...analysis. Our framework uses recurrence plots and recurrence quantification analysis to characterize dynamic networks. For resting‐state magnetoencephalographic dynamic functional networks (dFNs), we have found that functional networks recur more quickly in people with epilepsy than in healthy controls. This suggests that recurrence of dFNs may be used as a biomarker of epilepsy. For stereo electroencephalography data, we have found that dFNs involved in epileptic seizures emerge before seizure onset, and recurrence analysis allows us to detect seizures. We further observe distinct dFNs before and after seizures, which may inform neurostimulation strategies to prevent seizures. Our framework can also be used for understanding dFNs in healthy brain function and in other neurological disorders besides epilepsy.
We propose a new framework to assess the dynamics of brain networks based on recurrence analysis, which we applied to magnetoencephalographic (MEG) and stereo electroencephalographic (sEEG) recordings. We found that MEG functional networks recurred more quickly in people with epilepsy than in healthy controls. We further observed that sEEG dynamic functional networks involved in epileptic seizures emerged before seizure onset, and recurrence analysis allows us to detect seizures
Background
High-grade gliomas are among the most aggressive central nervous system primary tumors, with a high risk of recurrence and a poor prognosis. Re-operation, re-irradiation, chemotherapy are ...options in this setting. No-best therapy has been established. Bevacizumab was approved on the basis of two Phase 2 trials that evaluated its efficacy in patients with recurrent glioblastoma.
Materials and methods
We have retrospectively review data of patients with high-grade glioma treated at our institution that undergone radiological or histological progression after at least one systemic treatment for recurrent disease. Bevacizumab was administered alone or in combination with chemotherapy until disease progression or unacceptable toxicity. Bevacizumab regimen was analyzed to assess PFS and OS. Histological, molecular and clinical features of the entire cohort were collected.
Results
We reviewed data from 92 patients, treated from April 2009 to November 2019, with histologically confirmed diagnosis of high-grade gliomas and recurrent disease. A PFS of 55.2%, 22.9% and 9.6% was observed at 6, 12 and 24 months, respectively. Performance status, age at diagnosis (< 65 or > 65 ys.) and use of corticosteroids during bevacizumab therapy were strongly associated with PFS. The OS was 74.9% at 6 months, 31.7% at 12 months, 10.1% at 24 months. In our cohort, 51.1% were long-term responders (PFS > 6 months). Globally, bevacizumab treatment was well tolerated.
Conclusion
Our analysis confirms the efficacy of bevacizumab in recurrent high-grade glioma patients with an acceptable toxicity profile, in keeping with its known safety in the literature.
To evaluate a retrospective single-institution outcome after hypofractionated stereotactic body radiotherapy (SBRT) for adrenal metastases.
Between February 2002 and December 2009, we treated 48 ...patients with SBRT for adrenal metastases. The median age of the patient population was 62.7 years (range, 43-77 years). In the majority of patients, the prescription dose was 36 Gy in 3 fractions (70% isodose, 17.14 Gy per fraction at the isocenter). Eight patients were treated with single-fraction stereotactic radiosurgery and forty patients with multi-fraction stereotactic radiotherapy.
Overall, the series of patients was followed up for a median of 16.2 months (range, 3-63 months). At the time of analysis, 20 patients were alive and 28 patients were dead. The 1- and 2-year actuarial overall survival rates were 39.7% and 14.5%, respectively. We recorded 48 distant failures and 2 local failures, with a median interval to local failure of 4.9 months. The actuarial 1-year disease control rate was 9%; the actuarial 1- and 2-year local control rate was 90%.
Our retrospective study indicated that SBRT for the treatment of adrenal metastases represents a safe and effective option with a control rate of 90% at 2 years.