Visual attribution using Adversarial Latent Transformations

The ability to accurately locate all indicators of disease within medical images is vital for comprehending the effects of the disease, as well as for weakly-supervised segmentation and localization of the diagnostic correlators of disease. Existing methods either use classifiers to make predictions based on class-salient regions or else use adversarial learning based image-to-image translation to capture such disease effects. However, the former does not capture all relevant features for visual attribution (VA) and are prone to data biases; the latter can generate adversarial (misleading) and inefficient solutions when dealing in pixel values. To address this issue, we propose a novel approach Visual Attribution using Adversarial Latent Transformations (VA2LT). Our method uses adversarial learning to generate counterfactual (CF) normal images from abnormal images by finding and modifying discrepancies in the latent space. We use cycle consistency between the query and CF latent representations to guide our training. We evaluate our method on three datasets including a synthetic dataset, the Alzheimer's Disease Neuroimaging Initiative dataset, and the BraTS dataset. Our method outperforms baseline and related methods on all datasets.

Pediatrics in Artificial Intelligence Era: A Systematic Review on Challenges, Opportunities, and Explainability

BACKGROUND

The emergence of artificial intelligence (AI) tools such as ChatGPT and Bard is disrupting a broad swathe of fields, including medicine. In pediatric medicine, AI is also increasingly being used across multiple subspecialties. However, the practical application of AI still faces a number of key challenges. Consequently, there is a requirement for a concise overview of the roles of AI across the multiple domains of pediatric medicine, which the current study seeks to address.

AIM

To systematically assess the challenges, opportunities, and explainability of AI in pediatric medicine.

METHODOLOGY

A systematic search was carried out on peer-reviewed databases, PubMed Central, Europe PubMed Central, and grey literature using search terms related to machine learning (ML) and AI for the years 2016 to 2022 in the English language. A total of 210 articles were retrieved that were screened with PRISMA for abstract, year, language, context, and proximal relevance to research aims. A thematic analysis was carried out to extract findings from the included studies.

RESULTS

Twenty articles were selected for data abstraction and analysis, with three consistent themes emerging from these articles. In particular, eleven articles address the current state-of-the-art application of AI in diagnosing and predicting health conditions such as behavioral and mental health, cancer, syndromic and metabolic diseases. Five articles highlight the specific challenges of AI deployment in pediatric medicines: data security, handling, authentication, and validation. Four articles set out future opportunities for AI to be adapted: the incorporation of Big Data, cloud computing, precision medicine, and clinical decision support systems. These studies collectively critically evaluate the potential of AI in overcoming current barriers to adoption.

CONCLUSION

AI is proving disruptive within pediatric medicine and is presently associated with challenges, opportunities, and the need for explainability. AI should be viewed as a tool to enhance and support clinical decision-making rather than a substitute for human judgement and expertise. Future research should consequently focus on obtaining comprehensive data to ensure the generalizability of research findings.

Machine Learning in Pediatrics: Evaluating Challenges, Opportunities, and Explainability

BACKGROUND

The emergence of Artificial Intelligence (AI) tools such as ChatGPT and Bard is disrupting a broad swathe of fields, including medicine. In pediatric medicine, AI is also increasingly being used across multiple subspecialties. However, the practical application of AI still faces a number of key challenges. Consequently, there is a requirement for a concise overview of the roles of AI across the multiple domains of pediatric medicine that the current study seeks to address.

AIM

To systematically assess the challenges, opportunities, and explainability of AI in pediatric medicine.

METHODOLOGY

A systematic search was carried out on peer-reviewed databases, PubMed Central, Europe PubMed Central, and grey literature using search terms related to machine learning (ML) and AI for the years 2016 to 2022 in the English language. A total of 210 articles were retrieved that were screened with PRISMA for abstract, year, language, context, and proximal relevance to research aims. A thematic analysis was carried out to extract findings from the included studies.

RESULTS

Twenty articles were selected for data abstraction and analysis, with three consistent themes emerging from these articles. In particular, eleven articles address the current state-of-the-art application of AI in diagnosing and predicting health conditions such as behavioral and mental health, cancer, syndromic and metabolic diseases. Five articles highlight the specific challenges of AI deployment in pediatric medicines: data security, handling, authentication, and validation. Four articles set out future opportunities for AI to be adapted: the incorporation of Big Data, cloud computing, precision medicine, and clinical decision support systems. These studies collectively critically evaluate the potential of AI in overcoming current barriers to adoption.

CONCLUSION

AI is proving disruptive within pediatric medicine and is presently associated with challenges, opportunities, and the need for explainability. AI should be viewed as a tool to enhance and support clinical decision-making rather than a substitute for human judgement and expertise. Future research should consequently focus on obtaining comprehensive data to ensure the generalizability of research findings.

A user-guided personalization methodology to facilitate new smart home occupancy

Smart homes are becoming increasingly popular in providing people with the services they desire. Activity recognition is a fundamental task to provide personalised home facilities. Many promising approaches are being used for activity recognition; one of them is data-driven. It has some fascinating features and advantages. However, there are drawbacks such as the lack of ability to providing home automation from the day one due to the limited data available. In this paper, we propose an approach, called READY (useR-guided nEw smart home ADaptation sYstem) for developing a personalised automation system that provides the user with smart home services the moment they move into their new house. The system development process was strongly user-centred, involving users in every step of the system's design. Later, the user-guided transfer learning approach was introduced that uses an old smart home data set to enhance the existing smart home service with user contributions. Finally, the proposed approach and designed system were tested and validated in the smart lab that showed promising results.

Fully-Automated Identification of Imaging Biomarkers for Post-Operative Cerebellar Mutism Syndrome Using Longitudinal Paediatric MRI

Post-operative cerebellar mutism syndrome (POPCMS) in children is a post- surgical complication which occurs following the resection of tumors within the brain stem and cerebellum. High resolution brain magnetic resonance (MR) images acquired at multiple time points across a patient’s treatment allow the quantification of localized changes caused by the progression of this syndrome. However, MR images are not necessarily acquired at regular intervals throughout treatment and are often not volumetric. This restricts the analysis to 2D space and causes difficulty in intra- and inter-subject comparison. To address these challenges, we have developed an automated image processing and analysis pipeline. Multi-slice 2D MR image slices are interpolated in space and time to produce a 4D volumetric MR image dataset providing a longitudinal representation of the cerebellum and brain stem at specific time points across treatment. The deformations within the brain over time are represented using a novel metric known as the Jacobian of deformations determinant. This metric, together with the changing grey-level intensity of areas within the brain over time, are analyzed using machine learning techniques in order to identify biomarkers that correspond with the development of POPCMS following tumor resection. This study makes use of a fully automated approach which is not hypothesis-driven. As a result, we were able to automatically detect six potential biomarkers that are related to the development of POPCMS following tumor resection in the posterior fossa.

Representational fluidity in embodied (artificial) cognition

Theories of embodied cognition agree that the body plays some role in human cognition, but disagree on the precise nature of this role. While it is (together with the environment) fundamentally engrained in the so-called 4E (or multi-E) cognition stance, there also exists interpretations wherein the body is merely an input/output interface for cognitive processes that are entirely computational. In the present paper, we show that even if one takes such a strong computationalist position, the role of the body must be more than an interface to the world. To achieve human cognition, the computational mechanisms of a cognitive agent must be capable not only of appropriate reasoning over a given set of symbolic representations; they must in addition be capable of updating the representational framework itself (leading to the titular representational fluidity). We demonstrate this by considering the necessary properties that an artificial agent with these abilities need to possess. The core of the argument is that these updates must be falsifiable in the Popperian sense while simultaneously directing representational shifts in a direction that benefits the agent. We show that this is achieved by the progressive, bottom-up symbolic abstraction of low-level sensorimotor connections followed by top-down instantiation of testable perception-action hypotheses. We then discuss the fundamental limits of this representational updating capacity, concluding that only fully embodied learners exhibiting such a priori perception-action linkages are able to sufficiently ground spontaneously-generated symbolic representations and exhibit the full range of human cognitive capabilities. The present paper therefore has consequences both for the theoretical understanding of human cognition, and for the design of autonomous artificial agents.

Proteomic identification and characterization of hepatic glyoxalase 1 dysregulation in non-alcoholic fatty liver disease

Background

Non-alcoholic fatty liver disease (NAFLD) is the most common liver disease worldwide. However, its molecular pathogenesis is incompletely characterized and clinical biomarkers remain scarce. The aims of these experiments were to identify and characterize liver protein alterations in an animal model of early, diet-related, liver injury and to assess novel candidate biomarkers in NAFLD patients.

Methods

Liver membrane and cytosolic protein fractions from high fat fed apolipoprotein E knockout (ApoE^−/−) animals were analyzed by quantitative proteomics, utilizing isobaric tags for relative and absolute quantitation (iTRAQ) combined with nano-liquid chromatography and tandem mass spectrometry (nLC-MS/MS). Differential protein expression was confirmed independently by immunoblotting and immunohistochemistry in both murine tissue and biopsies from paediatric NAFLD patients. Candidate biomarkers were analyzed by enzyme-linked immunosorbent assay in serum from adult NAFLD patients.

Results

Through proteomic profiling, we identified decreased expression of hepatic glyoxalase 1 (GLO1) in a murine model. GLO1 protein expression was also found altered in tissue biopsies from paediatric NAFLD patients. In vitro experiments demonstrated that, in response to lipid loading in hepatocytes, GLO1 is first hyperacetylated then ubiquitinated and degraded, leading to an increase in reactive methylglyoxal. In a cohort of 59 biopsy-confirmed adult NAFLD patients, increased serum levels of the primary methylglyoxal-derived advanced glycation endproduct, hydroimidazolone (MG-H1) were significantly correlated with body mass index (r = 0.520, p < 0.0001).

Conclusion

Collectively these results demonstrate the dysregulation of GLO1 in NAFLD and implicate the acetylation-ubquitination degradation pathway as the functional mechanism. Further investigation of the role of GLO1 in the molecular pathogenesis of NAFLD is warranted.

Electronic supplementary material

The online version of this article (10.1186/s12953-018-0131-y) contains supplementary material, which is available to authorized users.

Movement correction in DCE-MRI through windowed and reconstruction dynamic mode decomposition

Images of the kidneys using dynamic contrast-enhanced magnetic resonance renography (DCE-MRR) contains unwanted complex organ motion due to respiration. This gives rise to motion artefacts that hinder the clinical assessment of kidney function. However, due to the rapid change in contrast agent within the DCE-MR image sequence, commonly used intensity-based image registration techniques are likely to fail. While semi-automated approaches involving human experts are a possible alternative, they pose significant drawbacks including inter-observer variability, and the bottleneck introduced through manual inspection of the multiplicity of images produced during a DCE-MRR study. To address this issue, we present a novel automated, registration-free movement correction approach based on windowed and reconstruction variants of dynamic mode decomposition (WR-DMD). Our proposed method is validated on ten different healthy volunteers' kidney DCE-MRI data sets. The results, using block-matching-block evaluation on the image sequence produced by WR-DMD, show the elimination of 99 % of mean motion magnitude when compared to the original data sets, thereby demonstrating the viability of automatic movement correction using WR-DMD.

Post-operative pediatric cerebellar mutism syndrome and its association with hypertrophic olivary degeneration

BACKGROUND

The dentato-thalamo-cortical (DTC) pathway is recognized as the anatomical substrate for postoperative pediatric cerebellar mutism (POPCMS), a well-recognized complication affecting up to 31% of children undergoing posterior fossa brain tumour resection. The proximal structures of the DTC pathway also form a segment of the Guillain and Mollaret triangle, a neural network which when disrupted causes hypertrophic olivary degeneration (HOD) of the inferior olivary nucleus (ION). We hypothesize that there is an association between the occurrence of POPCMS and HOD and aim to evaluate this on MR imaging using qualitative and quantitative analysis of the ION in children with and without POPCMS.

METHODS

In this retrospective study we qualitatively analysed the follow up MR imaging in 48 children who underwent posterior fossa tumour resection for presence of HOD. Quantitative analysis of the ION was possible in 28 children and was performed using semi-automated segmentation followed by feature extraction and feature selection techniques and relevance of the features to POPCMS were evaluated. The diagnosis of POPCMS was made independently based on clinical and nursing assessment notes.

RESULTS

There was significant association between POPCMS and bilateral HOD (P=0.002) but not unilateral HOD. Quantitative analysis showed that hyperintensity in the left ION was the most relevant feature in children with POPCMS.

CONCLUSIONS

Bilateral HOD can serve as a reliable radiological indicator in establishing the diagnosis of POPCMS particularly in equivocal cases. The strong association of signal change due to HOD in the left ION suggests that injury to the right proximal efferent cerebellar pathway plays an important role in the causation of POPCMS.

Classification of fibroglandular tissue distribution in the breast based on radiotherapy planning CT

BACKGROUND

Accurate segmentation of breast tissues is required for a number of applications such as model based deformable registration in breast radiotherapy. The accuracy of breast tissue segmentation is affected by the spatial distribution (or pattern) of fibroglandular tissue (FT). The goal of this study was to develop and evaluate texture features, determined from planning computed tomography (CT) data, to classify the spatial distribution of FT in the breast.

METHODS

Planning CT data of 23 patients were evaluated in this study. Texture features were derived from the radial glandular fraction (RGF), which described the distribution of FT within three breast regions (posterior, middle, and anterior). Using visual assessment, experts grouped patients according to FT spatial distribution: sparse or non-sparse. Differences in the features between the two groups were investigated using the Wilcoxon rank test. Classification performance of the features was evaluated for a range of support vector machine (SVM) classifiers.

RESULTS

Experts found eight patients and 15 patients had sparse and non-sparse spatial distribution of FT, respectively. A large proportion of features (>9 of 13) from the individual breast regions had significant differences (p <0.05) between the sparse and non-sparse group. The features from middle region had most significant differences and gave the highest classification accuracy for all the SVM kernels investigated. Overall, the features from middle breast region achieved highest accuracy (91%) with the linear SVM kernel.

CONCLUSION

This study found that features based on radial glandular fraction provide a means for discriminating between fibroglandular tissue distributions and could achieve a classification accuracy of 91%.

Identifying quantitative imaging features of posterior fossa syndrome in longitudinal MRI

Up to 25% of children who undergo brain tumor resection surgery in the posterior fossa develop posterior fossa syndrome (PFS). This syndrome is characterized by mutism and disturbance in speech. Our hypothesis is that there is a correlation between PFS and the occurrence of hypertrophic olivary degeneration (HOD) in structures within the posterior fossa, known as the inferior olivary nuclei (ION). HOD is exhibited as an increase in size and intensity of the ION on an MR image. Longitudinal MRI datasets of 28 patients were acquired consisting of pre-, intra-, and postoperative scans. A semiautomated segmentation process was used to segment the ION on each MR image. A full set of imaging features describing the first- and second-order statistics and size of the ION were extracted for each image. Feature selection techniques were used to identify the most relevant features among the MRI features, demographics, and data based on neuroradiological assessment. A support vector machine was used to analyze the discriminative features selected by a generative k-nearest neighbor algorithm. The results indicate the presence of hyperintensity in the left ION as the most diagnostically relevant feature, providing a statistically significant improvement in the classification of patients ([Formula: see text]) when using this feature alone.

Multilevel Chinese takeaway process and label-based processes for rule induction in the context of automated sports video annotation

We propose four variants of a novel hierarchical hidden Markov models strategy for rule induction in the context of automated sports video annotation including a multilevel Chinese takeaway process (MLCTP) based on the Chinese restaurant process and a novel Cartesian product label-based hierarchical bottom-up clustering (CLHBC) method that employs prior information contained within label structures. Our results show significant improvement by comparison against the flat Markov model: optimal performance is obtained using a hybrid method, which combines the MLCTP generated hierarchical topological structures with CLHBC generated event labels. We also show that the methods proposed are generalizable to other rule-based environments including human driving behavior and human actions.

Domain Anomaly Detection in Machine Perception: A System Architecture and Taxonomy

We address the problem of anomaly detection in machine perception. The concept of domain anomaly is introduced as distinct from the conventional notion of anomaly used in the literature. We propose a unified framework for anomaly detection which exposes the multifaceted nature of anomalies and suggest effective mechanisms for identifying and distinguishing each facet as instruments for domain anomaly detection. The framework draws on the Bayesian probabilistic reasoning apparatus which clearly defines concepts such as outlier, noise, distribution drift, novelty detection (object, object primitive), rare events, and unexpected events. Based on these concepts we provide a taxonomy of domain anomaly events. One of the mechanisms helping to pinpoint the nature of anomaly is based on detecting incongruence between contextual and noncontextual sensor(y) data interpretation. The proposed methodology has wide applicability. It underpins in a unified way the anomaly detection applications found in the literature. To illustrate some of its distinguishing features, in here the domain anomaly detection methodology is applied to the problem of anomaly detection for a video annotation system.

A Framework for Hierarchical Perception-Action Learning Utilizing Fuzzy Reasoning

Perception-action (P-A) learning is an approach to cognitive system building that seeks to reduce the complexity associated with conventional environment-representation/action-planning approaches. Instead, actions are directly mapped onto the perceptual transitions that they bring about, eliminating the need for intermediate representation and significantly reducing training requirements. We here set out a very general learning framework for cognitive systems in which online learning of the P-A mapping may be conducted within a symbolic processing context, so that complex contextual reasoning can influence the P-A mapping. In utilizing a variational calculus approach to define a suitable objective function, the P-A mapping can be treated as an online learning problem via gradient descent using partial derivatives. Our central theoretical result is to demonstrate top-down modulation of low-level perceptual confidences via the Jacobian of the higher levels of a subsumptive P-A hierarchy. Thus, the separation of the Jacobian as a multiplying factor between levels within the objective function naturally enables the integration of abstract symbolic manipulation in the form of fuzzy deductive logic into the P-A mapping learning. We experimentally demonstrate that the resulting framework achieves significantly better accuracy than using P-A learning without top-down modulation. We also demonstrate that it permits novel forms of context-dependent multilevel P-A mapping, applying the mechanism in the context of an intelligent driver assistance system.

Looking to score: the dissociation of goal influence on eye movement and meta-attentional allocation in a complex dynamic natural scene

Several studies have reported that task instructions influence eye-movement behavior during static image observation. In contrast, during dynamic scene observation we show that while the specificity of the goal of a task influences observers' beliefs about where they look, the goal does not in turn influence eye-movement patterns. In our study observers watched short video clips of a single tennis match and were asked to make subjective judgments about the allocation of visual attention to the items presented in the clip (e.g., ball, players, court lines, and umpire). However, before attending to the clips, observers were either told to simply watch clips (non-specific goal), or they were told to watch the clips with a view to judging which of the two tennis players was awarded the point (specific goal). The results of subjective reports suggest that observers believed that they allocated their attention more to goal-related items (e.g. court lines) if they performed the goal-specific task. However, we did not find the effect of goal specificity on major eye-movement parameters (i.e., saccadic amplitudes, inter-saccadic intervals, and gaze coherence). We conclude that the specificity of a task goal can alter observer's beliefs about their attention allocation strategy, but such task-driven meta-attentional modulation does not necessarily correlate with eye-movement behavior.

Perception-action learning as an epistemologically-consistent model for self-updating cognitive representation

As well as having the ability to formulate models of the world capable of experimental falsification, it is evident that human cognitive capability embraces some degree of representational plasticity, having the scope (at least in infancy) to modify the primitives in terms of which the world is delineated. We hence employ the term 'cognitive bootstrapping' to refer to the autonomous updating of an embodied agent's perceptual framework in response to the perceived requirements of the environment in such a way as to retain the ability to refine the environment model in a consistent fashion across perceptual changes.We will thus argue that the concept of cognitive bootstrapping is epistemically ill-founded unless there exists an a priori percept/motor interrelation capable of maintaining an empirical distinction between the various possibilities of perceptual categorization and the inherent uncertainties of environment modeling.As an instantiation of this idea, we shall specify a very general, logically-inductive model of perception-action learning capable of compact re-parameterization of the percept space. In consequence of the a priori percept/action coupling, the novel perceptual state transitions so generated always exist in bijective correlation with a set of novel action states, giving rise to the required empirical validation criterion for perceptual inferences. Environmental description is correspondingly accomplished in terms of progressively higher-level affordance conjectures which are likewise validated by exploratory action.Application of this mechanism within simulated perception-action environments indicates that, as well as significantly reducing the size and specificity of the a priori perceptual parameter-space, the method can significantly reduce the number of iterations required for accurate convergence of the world-model. It does so by virtue of the active learning characteristics implicit in the notion of cognitive bootstrapping.