Selected Publications by Michael Lones
IEEE Transactions on Evolutionary Computation, 2000
Biological organisms exist within environments in which complex, non-linear dynamics are ubiquito... more Biological organisms exist within environments in which complex, non-linear dynamics are ubiquitous. They are coupled to these environments via their own complex, dynamical networks of enzyme-mediated reactions, known as biochemical networks. These networks, in turn, control the growth and behaviour of an organism within its environment. In this paper, we consider computational models whose structure and function are motivated by the organisation of biochemical networks. We refer to these as artificial biochemical networks, and show how they can be evolved to control trajectories within three behaviourally diverse complex dynamical systems: the Lorenz system, Chirikov's standard map, and legged robot locomotion. More generally, we consider the notion of evolving dynamical systems to control dynamical systems, and discuss the advantages and disadvantages of using higher order coupling and configurable dynamical modules (in the form of discrete maps) within artificial biochemical networks. We find both approaches to be advantageous in certain situations, though note that the relative trade-offs between different models of artificial biochemical network strongly depend on the type of dynamical systems being controlled.
Lecture Notes in Computer Science, 2015
IEEE Transactions on Evolutionary Computation, 2000
Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion - GECCO Comp '14, 2014
In previous work, we have shown how an evolutionary algorithm with a clustered population can be ... more In previous work, we have shown how an evolutionary algorithm with a clustered population can be used to concurrently discover multiple regulatory motifs present within the promoter sequences of co-expressed genes. In this paper, we extend the algorithm by co-evolving a population of Boolean classification rules in parallel with the motif population. Results using synthetic data suggest that this approach allows poorly conserved motifs to be identified in promoter sequences a magnitude longer than using population clustering alone, whilst results using muscle-specific data suggest the algorithm is able to evolve meaningful sequence classifiers in parallel with motifs.
Papers by Michael Lones
Lecture Notes in Computer Science, 2015
2015 IEEE Congress on Evolutionary Computation (CEC), 2015
Natural Computing Series, 2015
IET Systems Biology, 2015
This study describes how the application of evolutionary algorithms (EAs) can be used to study mo... more This study describes how the application of evolutionary algorithms (EAs) can be used to study motor function in humans with Parkinson's disease (PD) and in animal models of PD. Human data is obtained using commercially available sensors via a range of non-invasive procedures that follow conventional clinical practice. EAs can then be used to classify human data for a range of uses, including diagnosis and disease monitoring. New results are presented that demonstrate how EAs can also be used to classify fruit flies with and without genetic mutations that cause Parkinson's by using measurements of the proboscis extension reflex. The case is made for a computational approach that can be applied across human and animal studies of PD and lays the way for evaluation of existing and new drug therapies in a truly objective way.
Lecture Notes in Computer Science, 2010
Artificial biochemical networks (ABNs) are computational models inspired by the biochemical netwo... more Artificial biochemical networks (ABNs) are computational models inspired by the biochemical networks which underlie the cellular activities of biological organisms. This paper shows how evolved ABNs may be used to control chaotic dynamics in both discrete and continuous dynamical systems, illustrating that ABNs can be used to represent complex computational behaviours within evolutionary algorithms. Our results also show that performance is sensitive to model choice, and suggest that conservation laws play an important role in guiding search.
Objective: To evaluate how accurately a new non-invasive device can monitor dyskinesia in people ... more Objective: To evaluate how accurately a new non-invasive device can monitor dyskinesia in people with Parkinson's. Background: Dyskinesia is a common and troublesome adverse e of drugs used to treat Parkinson's and is associated with reduced quality of life. It may be reduced by altering the drug regimen times and dosages. However the presence and severity of dyskinesia varies throughout the day, making it difficiult to monitor via outpatient clinics. It is costly and impractical to admit people for inpatient monitoring. A method of objectively monitoring dyskinesia over several hours in the patient's own home would inform drug regimen changes and improve management of dyskinesia. Methods: Ten people with Parkinson's dyskinesia were assessed for six hours on the neurology ward whilst wearing small movement sensors (Figures 1 & 2). Dyskinesia severity was graded clinically every hour according to the Unied Dyskinesia Rating Scale (UDysRS). The movement data from the s...
: a The movement of each sensor is measured in 3-dimensional space using 3 positional (x, y, z) a... more : a The movement of each sensor is measured in 3-dimensional space using 3 positional (x, y, z) and 3 orientation (azimuth, roll, elevation) coordinates; b two EM sensors attached to a subject's right hand.
ABSTRACT Objective: To assess whether an accurate objective measurement of bradykinesia can be ob... more ABSTRACT Objective: To assess whether an accurate objective measurement of bradykinesia can be obtained from novel devices employing electromagnetic (EM) tracking sensors and a digitising tablet. Background: Bradykinesia is the fundamental clinical feature of Parkinson's disease (PD) and may often be the sole motor presentation in the early stages. Clinical mis-diagnosis rates of PD may be as high as 20% even amongst experienced neurologists and not all clinicians have access to dopamine transporter imaging. A simple cheap non-invasive test is needed to provide objective measurement of bradykinesia in order to assist in accurate diagnosis and monitoring of PD. Methods: 49 PD patients and 41 age-matched controls produced drawings on a digitising tablet and performed finger-tapping tasks whilst wearing EM tracking sensors on the index finger and thumb. The mean PD diagnosis duration was 5.9 ±3.9 years (range 0.5-18) and mean H&Y stage was 2.5 ±0.7 (range 1-4). All patients were assessed whilst 'on' and the mean levodopa equivalent dose was 754 mg ±417 (range 140- 2080). The data recorded by the devices was analysed by custom-written computer evolutionary algorithms and compared to clinician rated scores using receiver operator characteristic (ROC) curves. Results: Clinical bradykinesia scores were normal in both hands for 4 (8%) patients and the rest were impaired in at least one hand as follows: 39 (80%) mild/moderate bradykinesia, 4 (8%) severe bradykinesia and 2 (4%) had such severe bradykinesia that they could barely perform the task. Classifiers induced from data recorded by the EM tracking sensor device had a median Area under ROC Curve (AUC) of 0.85 and those induced from the digitising tablet device had a median AUC of 0.84. The best classifiers had AUC scores approaching 0.9, corresponding to predictive accuracies of 80-90%, depending on the choice of threshold. Figure 1 shows ROC curves for a high scoring classifier with a circle marking the highest accuracy: sensitivity= 87.5% and specificity= 80%. Conclusions: This technology is able to provide an objective measurement of bradykinesia. Potential uses include aiding diagnosis of PD, monitoring response to treatment, evaluating new drugs and providing information about the more subtle characteristics of bradykinesia and hence underlying pathophysiology.
Lecture Notes in Computer Science, 2012
ABSTRACT Artificial gene regulatory networks are computational models which draw inspiration from... more ABSTRACT Artificial gene regulatory networks are computational models which draw inspiration from real world networks of biological gene regulation. Since their inception they have been used to infer knowledge about gene regulation and as methods of computation. These computational models have been shown to possess properties typically found in the biological world such as robustness and self organisation. Recently, it has become apparent that epigenetic mechanisms play an important role in gene regulation. This paper introduces a new model, the Artificial Epigenetic Regulatory Network (AERN) which builds upon existing models by adding an epigenetic control layer. The results demonstrate that the AERNs are more adept at controlling multiple opposing trajectories within Chirikov's standard map, suggesting that AERNs are an interesting area for further investigation.
Lecture Notes in Computer Science, 2012
ABSTRACT Artificial Signalling Networks (ASNs) are computational models inspired by cellular sign... more ABSTRACT Artificial Signalling Networks (ASNs) are computational models inspired by cellular signalling processes that interpret environmental information. This paper introduces an ASN-based approach to controlling chaotic dynamics in discrete dynamical systems, which are representative of complex behaviours which occur in the real world. Considering the main biological interpretations of signalling pathways, two ASN models are developed. They highlight how pathways' complex behavioural dynamics can be captured and represented within evolutionary algorithms. In addition, the regulatory capacity of the major regulatory functions within living organisms is also explored. The results highlight the importance of the representation to model signalling pathway behaviours and reveal that the inclusion of crosstalk positively affects the performance of the model.
2013 IEEE Congress on Evolutionary Computation, 2013
A novel bio-inspired architecture comprising three layers is introduced for a six-legged robot in... more A novel bio-inspired architecture comprising three layers is introduced for a six-legged robot in order to generate adaptive rhythmic locomotion patterns using environmental information. Taking inspiration from the intracellular signalling processes that decode environmental information, and considering the emergent behaviours that arise from the interaction of multiple signalling pathways, we develop a decentralised robot controller composed of a collection of artificial signalling networks. Crosstalk, a biological signalling mechanism, is used to couple such networks favouring their interaction. We also apply nonlinear oscillators to model gait generators, which induce symmetric and rhythmical locomotion movements. The trajectories are modulated by a coupled artificial signalling network, which yields adaptive and stable robotic locomotive patterns. Gait trajectories are converted into joint angles by means of inverse kinematics. The architecture is implemented in a simulated version of the real robot T-Hex. Our results demonstrate the ability of the architecture to generate adaptive and periodic gaits.
2013 IEEE International Conference on Evolvable Systems (ICES), 2013
In this paper we describe an Artificial Gene Regulatory Network (AGRN), whose form and function a... more In this paper we describe an Artificial Gene Regulatory Network (AGRN), whose form and function are inspired by biological epigenetics. This new architecture, termed an Artificial Epigenetic Network (AEN), is applied to the coupled inverted pendulum task, a control task that has complex non-linear dynamics. The AENs show significant benefits over previous AGRNs. Firstly, when applied to the coupled inverted pendulum task, they show a significant performance increase. In addition, the AENs self-partition, applying different genes to control different dynamics within the task, which is more analogous to gene regulation in nature. These networks also make it possible to gain user control over the dynamics of the network via the modification of the epigenetic layer.
Proceedings of the 2014 conference companion on Genetic and evolutionary computation companion - GECCO Comp '14, 2014
Parkinson's disease is a chronic neurodegenerative condition that manifests clinically with vario... more Parkinson's disease is a chronic neurodegenerative condition that manifests clinically with various movement disorders. These are often treated with the dopamine-replacement drug levodopa. However, the dosage of levodopa must be kept as low as possible in order to avoid the drug's side e↵ects, such as the involuntary, and often violent, muscle spasms called dyskinesia, or levodopa-induced dyskinesia. In this paper, we investigate the use of genetic programming for training classifiers that can monitor the e↵ectiveness of levodopa therapy. In particular, we evolve classifiers that can recognise tremor and dyskinesia, movement states that are indicative of insu cient or excessive doses of levodopa, respectively. The evolved classifiers achieve clinically useful rates of discrimination, with AUC>0.9. We also find that temporal classifiers generally out-perform spectral classifiers. By using classifiers that respond to low-level features of the data, we identify the conserved patterns of movement that are used as a basis for classification, showing how this approach can be used to characterise as well as classify abnormal movement.
Uploads
Selected Publications by Michael Lones
Papers by Michael Lones