Probabilistic Pattern Recognition Research Articles

Electrocardiographic changes predate Parkinson\u2019s disease onset

Autonomic nervous system involvement precedes the motor features of Parkinson’s disease (PD). Our goal was to develop a proof-of-concept model for identifying subjects at high risk of developing PD by analysis of cardiac electrical activity. We used standard 10-s electrocardiogram (ECG) recordings of 60 subjects from the Honolulu Asia Aging Study including 10 with prevalent PD, 25 with prodromal PD, and 25 controls who never developed PD. Various methods were implemented to extract features from ECGs including simple heart rate variability (HRV) metrics, commonly used signal processing methods, and a Probabilistic Symbolic Pattern Recognition (PSPR) method. Extracted features were analyzed via stepwise logistic regression to distinguish between prodromal cases and controls. Stepwise logistic regression selected four features from PSPR as predictors of PD. The final regression model built on the entire dataset provided an area under receiver operating characteristics curve (AUC) with 95% confidence interval of 0.90 [0.80, 0.99]. The five-fold cross-validation process produced an average AUC of 0.835 [0.831, 0.839]. We conclude that cardiac electrical activity provides important information about the likelihood of future PD not captured by classical HRV metrics. Machine learning applied to ECGs may help identify subjects at high risk of having prodromal PD.

A Study for Visual Realism of Designed Pictures on Computer Screens by Investigation and Brain-Wave Analyses.

In this article, the visual realism of designed pictures on computer screens is studied by investigation and brain-wave analyses. The practical electroencephalogram (EEG) measurement is always time-varying and fluctuating so that conventional statistical techniques are not adequate for analyses. This study proposes a new scheme based on "fingerprinting" to analyze the EEG. Fingerprinting is a technique of probabilistic pattern recognition used in electrical engineering, very like the identification of human fingerprinting in a criminal investigation. The goal of this study was to assess whether subjective preference for pictures could be manifested physiologically by EEG fingerprinting analyses. The most important advantage of the fingerprinting technique is that it does not require accurate measurement. Instead, it uses probabilistic classification. Participants' preference for pictures can be assessed using fingerprinting analyses of physiological EEG measurements.

Automated, high accuracy classification of Parkinsonian disorders: a pattern recognition approach.

Progressive supranuclear palsy (PSP), multiple system atrophy (MSA) and idiopathic Parkinson’s disease (IPD) can be clinically indistinguishable, especially in the early stages, despite distinct patterns of molecular pathology. Structural neuroimaging holds promise for providing objective biomarkers for discriminating these diseases at the single subject level but all studies to date have reported incomplete separation of disease groups. In this study, we employed multi-class pattern recognition to assess the value of anatomical patterns derived from a widely available structural neuroimaging sequence for automated classification of these disorders. To achieve this, 17 patients with PSP, 14 with IPD and 19 with MSA were scanned using structural MRI along with 19 healthy controls (HCs). An advanced probabilistic pattern recognition approach was employed to evaluate the diagnostic value of several pre-defined anatomical patterns for discriminating the disorders, including: (i) a subcortical motor network; (ii) each of its component regions and (iii) the whole brain. All disease groups could be discriminated simultaneously with high accuracy using the subcortical motor network. The region providing the most accurate predictions overall was the midbrain/brainstem, which discriminated all disease groups from one another and from HCs. The subcortical network also produced more accurate predictions than the whole brain and all of its constituent regions. PSP was accurately predicted from the midbrain/brainstem, cerebellum and all basal ganglia compartments; MSA from the midbrain/brainstem and cerebellum and IPD from the midbrain/brainstem only. This study demonstrates that automated analysis of structural MRI can accurately predict diagnosis in individual patients with Parkinsonian disorders, and identifies distinct patterns of regional atrophy particularly useful for this process.

Underwater positioning by kernel principal component analysis based probabilistic approach

In this paper, underwater acoustic positioning is given by kernel principal component analysis (KPCA) and maximum likelihood (ML). To reduce the impact of multi-path reflection on measured signals, we utilize location fingerprinting to implement positioning. In order to check whether the proposed positioning scheme has the ability to tolerate multi-path reflections or not, experiments are conducted in a confined towing tank with boundary walls. Different frequency components of a physical sound projector are viewed as the virtual sound projectors. Thus the required hardware is greatly reduced. Our positioning scheme is divided into two stages, which are offline training and online testing. In the training stage, underwater acoustic signals are collected at different pre-specified reference locations and then projected to the KPCA space. In the testing stage, underwater acoustic signals are collected at an unknown location and then projected to the KPCA space. Underwater positioning is given by probabilistic pattern recognition of maximum likelihood in the KPCA space. Finally, the Euclidean distance between the actual and estimated positions are calculated and taken as the positioning error. The results show that underwater positioning by KPCA based probabilistic approach is accurate and efficient.

A novel approach to probabilistic biomarker‐based classification using functional near‐infrared spectroscopy

Pattern recognition approaches to the analysis of neuroimaging data have brought new applications such as the classification of patients and healthy controls within reach. In our view, the reliance on expensive neuroimaging techniques which are not well tolerated by many patient groups and the inability of most current biomarker algorithms to accommodate information about prior class frequencies (such as a disorder's prevalence in the general population) are key factors limiting practical application. To overcome both limitations, we propose a probabilistic pattern recognition approach based on cheap and easy‐to‐use multi‐channel near‐infrared spectroscopy (fNIRS) measurements. We show the validity of our method by applying it to data from healthy controls (n = 14) enabling differentiation between the conditions of a visual checkerboard task. Second, we show that high‐accuracy single subject classification of patients with schizophrenia (n = 40) and healthy controls (n = 40) is possible based on temporal patterns of fNIRS data measured during a working memory task. For classification, we integrate spatial and temporal information at each channel to estimate overall classification accuracy. This yields an overall accuracy of 76% which is comparable to the highest ever achieved in biomarker‐based classification of patients with schizophrenia. In summary, the proposed algorithm in combination with fNIRS measurements enables the analysis of sub‐second, multivariate temporal patterns of BOLD responses and high‐accuracy predictions based on low‐cost, easy‐to‐use fNIRS patterns. In addition, our approach can easily compensate for variable class priors, which is highly advantageous in making predictions in a wide range of clinical neuroimaging applications. Hum Brain Mapp, 2013. © 2012 Wiley Periodicals, Inc.

Underwater acoustic localization by principal components analyses based probabilistic approach

In this paper, the underwater localization is given from wireless acoustic communication signals by probabilistic pattern recognition in eigenspace of PCA (principal components analyses). It should be emphasized that our underwater localization is from existing wireless acoustic communication signals, but not from additional localization systems. Our underwater localization scheme is based on fingerprinting and contains two stages, i.e., the off-line (i.e., training) and on-line (i.e., predicting) stages. In general, the received acoustic signals fluctuate seriously in underwater environments. To reduce the complexity and noise effects, all received signals are projected onto the eigenspace of PCA. Each projected feature is assumed to have Gaussian probabilistic distributions. Therefore, the location information can be easily obtained by probabilistic pattern recognition of projected features in PCA space. Note that our underwater localization scheme is not affected by reflected signals. To illustrate such a benefit, experiments were conducted in a bounded water pool where reflected signals exist near the walls. Experimental results show that the proposed underwater localization scheme is efficient and accurate. The proposed localization scheme is useful for underwater acoustic communication networks, and then in underwater technologies.

Uncertainties in Seismic Facies Analysis for Reservoir Characterisation or Monitoring: Causes and Consequences

Multivariate probabilistic pattern recognition methods such as discriminant analysis are powerful for interpreting seismic data in terms of reservoir properties. They consist of calibrating a mathematical relationship between local seismic information (seismic voxels or pieces of traces extracted at a reservoir level), characterised by a set of measurements, named seismic attributes, and a pre-defined set of classes, named seismic facies. In the context of seismic reservoir characterisation, this methodology provides a map or a volume of most probable facies, which describes the main reservoir heterogeneities. During the last decade, another application of this technique to the interpretation of 4D seismic data has also emerged: in this case, each seismic vintage is interpreted in terms of seismic facies, and differences in the interpretation of one vintage from another are correlated a posteriori to production effects such as fluid substitutions or pressure variations. As a probabilistic method, discriminant analysis allows the evaluation, at least partly, of uncertainty associated with seismic facies interpretation, and assignment probabilities computed by the methodology are a measure of the confidence of the interpretation. However, this uncertainty evaluation is only partial. In this paper, we will describe three other aspects of uncertainties, which are not correctly handled by classical discriminant analysis. The first aspect is linked with measurement errors. To propagate this type of error in seismic facies analysis, we have had to develop a methodology based on a combination of discriminant analysis and interval analysis. The other two aspects of uncertainty are linked with the interpretation of seismic facies. Due to the difference in resolution between seismic data in real reservoir heterogeneities, one seismic facies may correspond to several different fine-scale geological or dynamic scenarios. We will show in this paper how pseudo-well methodology, as well as seismic modelling (in the context of 4D interpretation) can help address this issue.

Anonymous Vehicle Tracking for Real-Time Traffic Surveillance and Performance on Signalized Arterials

One of the fundamental requirements for facilitating implementation of any advanced transportation management and information system (ATMIS) is the development of a real-time traffic surveillance system able to produce reliable and accurate traffic performance measures. This study presents a new framework for anonymous vehicle tracking capable of tracing individual vehicles by the vehicle features. The core part of the proposed vehicle tracking method is a vehicle reidentification algorithm for signalized intersections based on inductive vehicle signatures. The new vehicle reidentification system consists of two major components: search space reduction and probabilistic pattern recognition. Not only real-time intersection performance but also intersection origin–destination information can be obtained as the algorithm’s basic output. A systematic simulation investigation was conducted of the performance and feasibility of anonymous vehicle tracking on signalized arterials using the Paramics simulation model. Extensive research experience with vehicle reidentification techniques on single roadway segments was the basis for investigating the performance that could be obtained from tracking individual vehicles across multiple detector stations. The findings of this study serve as a logical and necessary precursor to possible field implementation of vehicle reidentification techniques. The proposed anonymous vehicle tracking methodology with existing traffic surveillance infrastructure would be an invaluable tool for operating agencies in support of ATMIS strategies for congestion monitoring, adaptive traffic control, system evaluation, and provision of real-time traveler information.

A central spectrum model: a synthesis of auditory-nerve timing and place cues in monaural communication of frequency spectrum.

A probabilistic psychophysical model for monaural communication from the auditory nerve to the brain is given in the form of a tonotopic display of stimulus spectrum, termed central spectrum. The model builds upon prior research demonstrating the potential of neural timing cues from the auditory nerve for conveying information on complex spectra, and was designed to meet the quantified demands of the psychophysics of frequency measurement. The central spectrum magnitude at each frequency is determined by the response of the auditory-nerve fibre with characteristic frequency matching that frequency. An interval histogram from each fiber is passed through a filter matched to the characteristic frequency of the fiber. This output versus characteristic frequency defines the central spectrum. Detailed analysis demonstrates that efficient probabilistic processing of the central spectrum described known psychophysical properties of frequency measurement in discrimination and periodicity pitch experiments. Psychophysical models based upon the central spectrum model followed by optimum probabilistic pattern recognition are potentially relevant for predicting human communication limits in response to arbitrary sounds of speech and music.

Alternative k-nearest neighbour rules in supervised pattern recognition : Part 2. Probabilistic classification on the basis of the kNN method modified for direct density estimation

An implementation of the kNN classification method of Loftsgaarden and Quesenberry is discussed. It is a theoretically valid alternative for probabilistic pattern recognition by means of the Bayes equation. The method is illustrated on the basis of a data set containing laboratory tests for three categories of functional states of the thyroid gland and the performance is compared with those of other probabilistic pattern recognition techniques. Although the results are acceptable, the kNN method did not perform as well as some other probabilistic techniques.

Alternative k-nearest neighbour rules in supervised pattern recognition : Part 3. Condensed nearest neighbour rules

Two nearest neighbour rules in which the training set is reduced to a limited number of representative objects (condensed nearest neighbour method) are discussed. The emphasis is on the probabilistic approach. The performance criteria of probabilistic methods are discussed and the performance of the condensed nearest neighbour method is compared with that of other probabilistic pattern recognition methods. This is done on the basis of a data set concerning the differentiation of thyroid functional states.