Artificial Neural Network Output Research Articles

Artificial neural networks in tandem with molecular descriptors as predictive tools for continuous liposome manufacturing

The current study utilized an artificial neural network (ANN) to generate computational models to achieve process optimization for a previously developed continuous liposome manufacturing system. The liposome formation was based on a continuous manufacturing system with a co-axial turbulent jet in a co-flow technology. The ethanol phase with lipids and aqueous phase resulted in liposomes of homogeneous sizes. The input features of the ANN included critical material attributes (CMAs) (e.g., hydrocarbon tail length, cholesterol percent, and buffer type) and critical process parameters (CPPs) (e.g., solvent temperature and flow rate), while the ANN outputs included critical quality attributes (CQAs) of liposomes (i.e., particle size and polydispersity index (PDI)). Two common ANN architectures, multiple-input-multiple-output (MIMO) models and multiple-input–single-output (MISO) models, were evaluated in this study, where the MISO outperformed MIMO with improved accuracy. Molecular descriptors, obtained from PaDEL-Descriptor software, were used to capture the physicochemical properties of the lipids and used in training of the ANN. The combination of CMAs, CPPs, and molecular descriptors as inputs to the MISO ANN model reduced the training and testing mean relative error. Additionally, a graphic user interface (GUI) was successfully developed to assist the end-user in performing interactive simulated risk analysis and visualizing model predictions.

Performances of Artificial Neural Network (ANN) and Particle Swarm Optimization (PSO) Using KDD Cup ‘99 Dataset in Intrusion Detection System (IDS)

Nowadays, the number of attacker increasing fast due to the current technologies. Most companies or an organization use Intrusion Detection System (IDS) to protect their network system. Many researchers suggest different ways to improve the IDS such using optimization techniques. Artificial Intelligence (AI) methods also proposed in IDS to attained high accuracy of detection for example; artificial neural network, particle swarm optimization and genetic algorithm. Artificial neural network (ANN) and Particle Swarm Optimization (PSO) used in this paper to equate the method and performances in IDS environment. The ANN output value will be compared with the result where ANN supported by PSO to produce higher accurate value. KDD CUP ’99 Dataset used as the benchmark of IDS and will be simulated in MATLAB Simulink 2013. 200 datasets used consists of attacks and normal activities as an input. In this paper, DoS attack which are Smurf and Neptune attacks selected for detection.

Numerical Simulation and Hydrodynamic Performance Predicting of 2 Two-Dimensional Hydrofoils in Tandem Configuration

In this study we investigated the performance of NACA 0012 hydrofoils aligned in tandem using parametric method and Neural Networks. We use the 2D viscous numerical model (STAR-CCM+) to simulate the hydrofoil system. To validate the numerical model, we modeled a single NACA 0012 configuration and compared it to experimental results. Results are found in concordance with the published experimental results. Then two NACA 0012 hydrofoils in tandem configuration were studied in relation to 788 combinations of the following parameters: spacing between two hydrofoils, angle of attack (AOA) of upstream hydrofoil and AOA of downstream hydrofoil. The effects exerted by these three parameters on the hydrodynamic coefficients Lift coefficient (CL), Drag Coefficient (CD) and Lift-Drag Ratio (LDR), are consistent with the behavior of the system. To establish a control system for the hydrofoil craft, a timely analysis of the hydrodynamic system is needed due to the computational resource constraints, analysis of a large combination and time consuming of the three parameters established. To provide a broader and faster way to predict the hydrodynamic performance of two hydrofoils in tandem configuration, an optimal artificial neural network (ANN) was trained using the large combination of three parameters generated from the numerical simulations. Regression analysis of the output of ANN was performed, and the results are consistent with numerical simulation with a correlation coefficient greater than 99.99%. The optimized spacing of 6.6c are suggested where the system has the lowest CD while obtaining the highest CL and LDR. The formula of the ANN was then presented, providing a reliable predicting method of hydrofoils in tandem configuration.

Designing an Intelligent Control Philosophy in Reservoirs of Water Transfer Networks in Supervisory Control and Data Acquisition System Stations

In this paper, a hybrid neural-genetic fuzzy system is proposed to control the flow and height of water in the reservoirs of water transfer networks. These controls will avoid probable water wastes in the reservoirs and pressure drops in water distribution networks. The proposed approach combines the artificial neural network, genetic algorithm, and fuzzy inference system to improve the performance of the supervisory control and data acquisition stations through a new control philosophy for instruments and control valves in the reservoirs of the water transfer networks. First, a multi-core artificial neural network model, including a multi-layer perceptron and radial based function, is proposed to forecast the daily consumption of the water in a reservoir. A genetic algorithm is proposed to optimize the parameters of the artificial neural networks. Then, the online height of water in the reservoir and the output of artificial neural networks are used as inputs of a fuzzy inference system to estimate the flow rate of the reservoir inlet. Finally, the estimated inlet flow is translated into the input valve position using a transform control unit supported by a nonlinear autoregressive exogenous model. The proposed approach is applied in the Tehran water transfer network. The results of this study show that the usage of the proposed approach significantly reduces the deviation of the reservoir height from the desired levels.

Artificial neural network approach for solving fuzzy fractional order initial value problems under gH‐differentiability

This paper aims to solve the celebrated fuzzy fractional differential equations (FFDEs) using an artificial neural network (ANN) technique. To accomplish the aforementioned aim, the error back propagation algorithm and a multilayer feed forward neural architecture are utilized using the unsupervised learning in order to minimize the error function as well as the modification of the parameters such as weights and biases. By combining the initial conditions with the ANN output provides an appropriate approximate solution of the proposed FFDE. Then, two illustrative examples are solved to confirm the applicability of the concept as well as to demonstrate both the precision and effectiveness of the developed method. By comparing with some traditional methods, the obtained results reveal a close match that confirms both accuracy and correctness of the proposed method.

Application of Artificial Neural Network to Somatotype Determination

Somatotype characteristics are important for the selection of sporting activities, as well as and the prevalence of several chronic diseases. Nowadays the most common method of somatotyping is the Heath–Carter method, which calculates the somatotype base on 10 anthropometric parameters. Another possibility for evaluation of somatotype gives commonly used bioelectrical impedance analysis), but the accuracy of the proposed formulas is questioned. Therefore, we aimed to investigate the possibility of applying an artificial neural network to achieve the formulas, which allow us to determine the endomorphy and mesomorphy using data on body height and weight and raw bioelectrical impedance analysis data in young women. The endomorphy (Endo), ectomorphy (Ecto), and mesomorphy (Meso) ratings were determined using artificial neural networks and the Heath–Carter method. To identify critical parameters and their degree of impact on the artificial neural network outputs, a sensitivity analysis was performed. The multi-layer perceptron MLP 4-4-1 (input: body mass index (BMI), reactance, resistance, and resting metabolic rate) for the Endo somatotype was proposed (root mean squared error (RMSE) = 0.66, χ2 = 0.66). The MLP 4-4-1 (input: BMI, fat-free mass, resistance, and total body water) for the Meso somatotype was proposed (RMSE = 0.76, χ2 = 0.87). All somatotypes (Endo, Meso and Ecto) can be calculated using MLP 2-4-3 (input: BMI and resistance) with accuracy RMSE = 0.67 and χ2 = 0.51. The bioelectrical impedance analysis and Heath–Carter method compliance was evaluated with the statistical algorithm proposed by Bland and Altman. The artificial neural network-based formulas allow us to determine the endomorphy and mesomorphy in young women’s ratings with high accuracy and agreement with the Heath–Carter method. The results of our study indicate the successful application of artificial neural network-based model in predicting the somatotype of young women. The artificial neural network model can be practically used in bioelectrical impedance analysis devices in the future.

Production of Yttrium-86 radioisotope using genetic algorithm and neural network

Recently, there has been a great deal of attention for applying radioisotopes in medical applications such as photon or positron emission tomography. In this study, TALYS code was utilized for prediction of nuclear reactions and led to a cross section calculation along with excitation function assessment of different nuclear reactions for production of target radioisotopes in medical areas. Subsequently, some parameters related to this code were changed in different reactions to achieve optimal outputs. Here, the range of optimal proton energy from 12–16 MeV, 86Sr target thickness and the 86Y production gain were calculated. The obtained data were optimized by using Artificial Neural Network (ANN) and Genetic Algorithm (GA). At 14 MeV, ANN revealed the greater generated 86Y while reducing contamination against trained bad neural network in wrong conditions. Also through GA optimization method and using ANN outputs, the average error achieved a 15 % improvement in ANN performance versus wrong training outputs. The rate of produced pollutant decreased significantly and the errors diminished due to the proper training of GA performance. This procedure may affect the control of radioactive contamination in medical radioisotope production to decrease the imposed absorbed dose on the patient.

Artificial Neural Network-Based Compact Modeling Methodology for Advanced Transistors

The artificial neural network (ANN)-based compact modeling methodology is evaluated in the context of advanced field-effect transistor (FET) modeling for Design-Technology-Cooptimization (DTCO) and pathfinding activities. An ANN model architecture for FETs is introduced, and the results clearly show that by carefully choosing the conversion functions (i.e., from ANN outputs to device terminal currents or charges) and the loss functions for ANN training, ANN models can reproduce the current-voltage and charge-voltage characteristics of advanced FETs with excellent accuracy. A few key techniques are introduced in this work to enhance the capabilities of ANN models (e.g., model retargeting, variability modeling) and to improve ANN training efficiency and SPICE simulation turn-around-time (TAT). A systematical study on the impact of the ANN size on ANN model accuracy and SPICE simulation TAT is conducted, and an automated flow for generating optimum ANN models is proposed. The findings in this work suggest that the ANN-based methodology can be a promising compact modeling solution for advanced DTCO and pathfinding activities.

Hydrodynamic Performance Prediction of Stepped Planing Craft Using CFD and ANNs

In the present paper, the hydrodynamic performance of stepped planing craft is investigated by computational fluid dynamics (CFD) analysis. For this purpose, the hydrodynamic resistances of without step, one-step, and two-step hulls of Cougar planing craft are evaluated under different distances of the second step and LCG from aft, weight loadings, and Froude numbers (Fr). Our CFD results are appropriately validated against our conducted experimental test in National Iranians Marine Laboratory (NIMALA), Tehran, Iran. Then, the hydrodynamic resistance of intended planing crafts under various geometrical and physical conditions is predicted using artificial neural networks (ANNs). CFD analysis shows two different trends in the growth rate of resistance to weight ratio. So that, using steps for planing craft increases the resistance to weight ratio at lower Fr and decreases it at higher Fr. Additionally, by the increase of the distance between two steps, the resistance to weight ratio is decreased and the porpoising phenomenon is delayed. Furthermore, we obtained the maximum mean square error of ANNs output in the prediction of resistance to weight ratio equal to 0.0027. Finally, the predictive equation is suggested for the resistance to weight ratio of stepped planing craft according to weights and bias of designed ANNs.

Artificial Intelligence-Based Protocol for Macroscopic Traffic Simulation Model Development

This study proposes a machine learning-based protocol for developing a TRANSYT-7F model for an urban arterial network. The developed artificial neural network (ANN) method models the queue lengths of TRANSYT-7F using saturation flow, start-up lost time, and platoon dispersion as inputs. The queue lengths of the selected approaches of the study network can be obtained using the ANN model without running the TRANSYT-7F model. The optimum values of the selected parameters (i.e., saturation flow, start-up lost time, and platoon dispersion) were obtained using the genetic algorithm, which ensures minimum difference between the measured queue length and the ANN output (i.e., queue length). Finally, the comparison of the measured queue length and the simulated queue length with the calibrated TRANSYT-7F model revealed that the mean absolute percentage error was less than 2.5% for all approaches of the study network.

Optimized ANN model for predicting rock mass quality ahead of tunnel face using measure-while-drilling data

Rock mass quality assessment has a vital influence on the excavation of tunnels and caverns in rock mass. For this purpose, extensive field studies, including records of measure-while-drilling data and rock mass quality scores (RQS) from the observation reports of tunnel faces, have been conducted. In order to predict RQS, three optimized artificial neural network (ANN) models based on genetic algorithm (GA), particle swarm optimization (PSO), and imperialist competition algorithm (ICA) were developed. Six parameters of measure-while-drilling (MWD) data and their corresponding RQS constituted 1270 datasets, which were set as input and output of ANN, respectively. The traditional multiple linear regression (MLR), multiple nonlinear regression (MNR) statistical model, and ANN model were developed as comparative models. Comparison results reveal that PSO-ANN and ICA-ANN models are capable of predicting RQS with higher reliability than the MLR, MNR, ANN, and GA-ANN models. Results indicate that PSO-ANN and ICA-ANN models can be used to predict RQS; however, the PSO-ANN model has better performance.

Response surface methodology and artificial neural network modelling for enhancing maturity parameters during vermicomposting of floral waste

In this study, the mixture of floral waste and cattle dung in different proportions was utilised to convert into vermicompost using earthworm Eisenia fetida. In the design of the experiment, the optimum amount of floral waste (1325 g) and cattle dung (500 g) was obtained for vermicompost using central composite design (CCD) and compared with the output of artificial neural network (ANN). The optimum proportions of vermicompost showed pH of 7.10, electrical conductivity of 3.39 mS/cm, total organic carbon of 34.01%, C: N ratio of 13, phosphorous of 5.31 g/kg and potassium of 14.45 g/kg. This vermicompost was enriched with sufficient concentration of nutrients like potassium, sodium, phosphorous, and calcium, which are beneficial for the growth of the plants. The current study was based on comparing response surface methodology (RSM) and ANN for maturity parameters and the value of R2 in both the cases was near 1.

A Damage Detection Method Using Neural Network Optimized by Multiple Particle Collision Algorithm

A critical task of structural health monitoring is damage detection and localization. Lamb wave propagation methods have been successfully applied for damage identification in plate‐like structures. However, Lamb wave processing is still a challenging task due to its multimodal and dispersive characteristics. To address this issue, data‐driven machine learning approaches as artificial neural network (ANN) have been proposed. However, the effectiveness of ANN can be improved based on its architecture and the learning strategy employed to train it. The present paper proposes a Multiple Particle Collision Algorithm (MPCA) to design an optimum ANN architecture to detect and locate damages in plate‐like structures. For the first time in the literature, the MPCA is applied to find damages in plate‐like structures. The present work uses one piezoelectric transducer to generate Lamb wave signals on an aluminum plate structure and a linear array of four transducers to capture the scattered signals. The continuous wavelet transform (CWT) processes the captured signals to estimate the time‐of‐flight (ToF) that is the ANN inputs. The ANN output is the damage spatial coordinates. In addition to MPCA optimization, this paper uses a quantitative entropy‐based criterion to find the best mother wavelet and the scale values. The presented experimental results show that MPCA is capable of finding a simple ANN architecture with good generalization performance in the proposed damage localization application. The proposed method is compared with the 1‐dimensional convolutional neural network (1D‐CNN). A discussion about the advantages and limitations of the proposed method is presented.

Short term wind energy resource prediction using WRF model for a location in western part of Turkey

Wind energy is a rapidly growing industry in Turkey. Wind power potential studies revealed that the most promising region for electricity generation is the western part of Turkey. Wind speed forecasting is necessary for power systems because of the intermittent nature of wind. Thus, accurate forecasting of wind power is recognized as a major contribution to reliable wind power integration. This paper assesses the performance of the weather research forecasting (WRF) model for wind speed and wind direction predictions up to 72 h ahead. The wind speeds and wind directions are evaluated based on the mean absolute error (MAE). Evaluations were also performed seasonally. Moreover, in order to improve the WRF simulations, a multi-input–single output artificial neural network (ANN) approach is applied to both wind speeds of the WRF model and wind power estimates, which are estimated from the wind speeds of the WRF model by using a power curve for the Soma wind power plant. Traditional error metrics were used for validations using wind tower mast data installed nearby the wind farm. The results from up to 72 h forecast horizon show that the WRF model slightly overpredicts the wind speeds. Wind speed predictions by the WRF model are found highly depending on the season, location, and wind direction. The model is also able to reproduce wind directions except for low wind speeds. Large MAEs are found for the winds less than 5 m/s. The performance of the WRF model for wind power prediction decreases with the increasing runtime. Root mean square error and normalized root mean square error (nRMSE) in wind powers range in between 123–261 kW and 13%–32% without performing the ANN approach, respectively. The improvement of the ANN depends on the forecast horizon, season, and location of turbine groups, as well as its application on either the wind speed outputs of the WRF model or wind power estimations. The ANN significantly improves the WRF at large forecast horizons for wind power estimations, for which it gives better results in the summer and reaches 29% improvement for summer on average for nRMSE. On the other hand, ANN adjusts the wind speed outputs of the model better than that of wind power estimations. For instance, the nRMSE is approximately 13% for 24 h winter wind speed simulations of the WRF for the turbine groups G1 and G4, after ANN adjustment. The ANN improves the results better for turbine group 1, because of less complexity of this group in the direction of prevailing wind. The evaluation of the ANN suggests that the approach can be used for improving the performance of the wind power forecast for this power plant.

Environmental impacts cost assessment model of residential building using an artificial neural network

PurposeBuildings are major contributors to greenhouse gases (GHG) along the various stages of the building life cycle. A range of tools have been utilised for estimating building energy use and environmental impacts; these are time-consuming and require massive data that are not necessarily available during early design stages. Therefore, this study aimed to develop an Environmental Impacts Cost Assessment Model (EICAM) that quantifies both energy and environmental costs for residential buildings.Design/methodology/approachAn Artificial Neural Network (ANN) was employed to develop the EICAM. The model consists of six input parameters, including wall type, roof type, glazing type, window to wall ratio (WWR), shading device and building orientation. In addition, the model calculates four measures: annual energy cost, operational carbon over 20 years, envelope embodied carbon and total carbon per square metre. The ANN architecture is 6:13:4:4, where the conjugate gradient algorithm was applied to train the model and minimise the mean squared error (MSE). Furthermore, regression analysis for the ANN prediction for each output was performed.FindingsThe MSE was minimised to 0.016 while training the model. Also, the correlation between each ANN output and the actual output was very strong, with an R2 value for each output of almost 0.998. Moreover, validation was conducted for each output, with the error percentages calculated at 0.26%, 0.25%, 0.03% and 0.27% for the annual energy cost, operational carbon, envelope materials embodied carbon and total carbon per square metre, respectively. Accordingly, the EICAM contributes to enhancing design decision-making concerning energy consumption and carbon emissions in the early design stages.Research limitations/implicationsThis study provides theoretical implications to the domain of building environmental impact assessment through illustrating a systematic approach for developing an energy-based prediction model that generates four environmental-oriented outputs, namely energy cost, operational energy carbon, envelope embodied carbon, and total carbon. The model developed has practical implications for the architectural/engineering (A/E) industries by providing a useful tool to easily predict environmental impact costs during the early design phase. This would enable designers in Saudi Arabia to make effective design decisions that would increase sustainability in the building life cycle.Originality/valueBy providing a holistic predictive model entitled EICAM, this study endeavours to bridge the gap between energy costs and environmental impacts in a predictive model for Saudi residential units. The novelty of this model is that it is an alternative tool that quantifies both energy cost, as well as building’s environmental impact, in one model by using a machine learning approach. Besides, EICAM predicts its outcomes more quickly than conventional tools such as DesignBuilder and is reliable for predicting accurate environmental impact costs during early design stages.

Multistep Flood Inundation Forecasts with Resilient Backpropagation Neural Networks: Kulmbach Case Study

Flooding, a significant natural disaster, attracts worldwide attention because of its high impact on communities and individuals and increasing trend due to climate change. A flood forecast system can minimize the impacts by predicting the flood hazard before it occurs. Artificial neural networks (ANN) could efficiently process large amounts of data and find relations that enable faster flood predictions. The aim of this study is to perform multistep forecasts for 1–5 h after the flooding event has been triggered by a forecast threshold value. In this work, an ANN developed for the real-time forecast of flood inundation with a high spatial resolution (4 m × 4 m) is extended to allow for multiple forecasts. After trained with 120 synthetic flood events, the ANN was first tested with 60 synthetic events for verifying the forecast performance for 3 h, 6 h, 9 h and 12 h lead time. The model produces good results, as shown by more than 81% of all grids having an RMSE below 0.3 m. The ANN is then applied to the three historical flood events to test the multistep inundation forecast. For the historical flood events, the results show that the ANN outputs have a good forecast accuracy of the water depths for (at least) the 3 h forecast with over 70% accuracy (RMSE within 0.3 m), and a moderate accuracy for the subsequent forecasts with (at least) 60% accuracy.

Artificial Neural Networks in the Prediction of Genetic Merit to Flowering Traits in Bean Cultivars

Flowering is an important agronomic trait that presents non-additive gene action. Genome-enabled prediction allow incorporating molecular information into the prediction of individual genetic merit. Artificial neural networks (ANN) recognize patterns of data and represent an alternative as a universal approximation of complex functions. In a Genomic Selection (GS) context, the ANN allows automatically to capture complicated factors such as epistasis and dominance. The objectives of this study were to predict the individual genetic merits of the traits associated with the flowering time in the common bean using the ANN approach, and to compare the predictive abilities obtained for ANN and Ridge Regression Best Linear Unbiased Predictor (RR-BLUP). We used a set of 80 bean cultivars and genotyping was performed with a set of 384 SNPs. The higher accuracy of the selective process of phenotypic values based on ANN output values resulted in a greater efficacy of the genomic estimated breeding value (GEBV). Through the root mean square error computational intelligence approaches via ANN, GEBV were shown to have greater efficacy than GS via RR-BLUP.

Application of gene expression programming, artificial neural network and multilinear regression in predicting hydrochar physicochemical properties

Globally, the provision of energy is becoming an absolute necessity. Biomass resources are abundant and have been described as a potential alternative source of energy. However, it is important to assess the fuel characteristics of the various available biomass sources. Soft computing techniques are presented in this study to predict the mass yield (MY), energy yield (EY), and higher heating value (HHV) of hydrothermally carbonized biomass using Gene Expression Programming (GEP), multiple-input single output-artificial neural network (MISO-ANN), and Multilinear regression (MLR). The three techniques were compared using statistical performance metrics. The coefficient of determination (R2), mean absolute error (MAE) and mean bias error (MBE) were used to evaluate the performance of the models. The MISO-ANN with 5-10 to 10-1 and 5-15-15-1 network architectures provided the most satisfactory performance of the three proposed models (R2 = 0.976, 0.955, 0.996; MAE = 2.24, 2.11, 0.93; MBE = 0.16, 0.37, 0.12) for MY, EY and HHV, respectively. The GEP technique’s ability to predict hydrochar properties based on the input parameters was found to be satisfactory, while MLR provided an unsatisfactory predictive model. Sensitivity analysis was conducted, and the analysis revealed that volatile matter (VM) and temperature (Temp) have more influence on the MY, EY, and HHV.

Predicting Estimated Blood Loss and Transfusions in Gynecologic Surgery Using Artificial Neural Networks

This chapter explores valuating the efficacy of using artificial neural networks (ANNs) for predicting the estimated blood loss (EBL) and also transfusion requirements of myomectomy patients. All 146 myomectomy surgeries performed over a 6-year period from a single site are captured. Records were removed for various reasons, leaving 96 cases. Backpropagation and radial basis function ANN models were developed to predict EBL and perioperative transfusion needs along with a regression model. The single hidden layer backpropagation ANN models performed the best for both prediction problems. EBL was predicted on average within 127.33 ml of measured blood loss, and transfusions were predicted with 71.4% sensitivity and 85.4% specificity. A combined ANN ensemble model using the output of the EBL ANN as an input variable to the transfusion prediction ANN was developed and resulted in 100% sensitivity and 62.9% specificity. The preoperative identification of large EBL or transfusion need can assist caregivers in better planning for possible post-operative morbidity and mortality.

Detection of transient neurotransmitter response using personalized neural networks

Measurement of stimulus-induced dopamine release and other types of transient neurotransmitter response (TNR) from dynamic positron emission tomography (PET) images typically suffers from limited detection sensitivity and high false positive (FP) rates. Measurement of TNR of a voxel-level can be particularly problematic due to high image noise. In this work, we perform voxel-level TNR detection using artificial neural networks (ANN) and compare their performance to previously used standard statistical tests. Different ANN architectures were trained and tested using simulated and real human PET imaging data, obtained with the tracer [11C]raclopride (a D2 receptor antagonist). A distinguishing feature of our approach is the use of ‘personalized’ ANNs that are designed to operate on the image from a specific subject and scan. Training of personalized ANNs was performed using simulated images that have been matched with the acquired image in terms of the signal, resolution, and noise.In our tests of TNR detection performance, the F-test of the linear parametric neurotransmitter PET model fit residuals was used as the reference method. For a moderate TNR magnitude, the areas under the receiver operating characteristic curves in simulated tests were 0.64 for the F-test and 0.77–0.79 for the best ANNs. At a fixed FP rate of 0.01, the true positive rates were 0.6 for the F-test and 0.8–0.9 for the ANNs. The F-test detected on average 28% of a 8.4 mm cluster with a strong TNR, while the best ANN detected 47%. When applied to a real image, no significant abnormalities in the ANN outputs were observed. These results demonstrate that personalized ANNs may offer a greater detection sensitivity of dopamine release and other types of TNR compared to previously used method based on the F-test.