Transfer Learning in Computer Vision for the Classification of Alzheimer’s Stages with MRI.

doi:10.22167/2675-6528-202602694

Article

September 30, 2026

Transfer Learning in Computer Vision for the Classification of Alzheimer’s Stages with MRI.

Transfer Learning in Computer Vision for the Classification of Alzheimer’s Stages with MRI.

André Pirini Spadaro; Gabriel Gomes de Oliveira

DOI: 10.22167/2675-6528-202602694

Article derived from a Course Conclusion Work (TCC), with content based on the student’s original work and adapted to the editorial format of the E&S Magazine with the support of the ResumeAI tool, an artificial intelligence solution developed by Instituto Pecege for textual synthesis and organization.

Summary

Alzheimer’s disease is one of the greatest global health challenges, with early and accurate diagnosis being crucial for patient management. A deep learning model was developed and validated for the automatic classification of the four stages of the disease, using 6400 brain magnetic resonance imaging (MRI) scans. Transfer learning methodology was employed, with the pre-trained Convolutional Neural Network (CNN) ResNet18 architecture. The model was trained in two phases, feature extraction followed by fine-tuning, on a public dataset divided into training, validation, and test sets, with no additional patient information. The results on the test set were promising, achieving an overall accuracy of 99.3% and a Cohen’s Kappa coefficient of 0.988. The model demonstrated remarkable generalization capability, perfectly identifying the rarest class, Moderate Alzheimer’s, with a precision and sensitivity of 1.000. It was concluded that the approach is highly effective and holds great potential as a support tool for the clinical diagnosis of Alzheimer’s disease, contributing to the improvement of patients’ quality of life through early detection.

Keywords: Deep Learning; Convolutional Neural Networks (CNN); ResNet18; Brain Magnetic Resonance Imaging (MRI); Transfer Learning.

1. Introduction

Alzheimer’s disease (AD) is recognized as the leading cause of major neurocognitive disorder on a global scale. With the progressive aging of the world’s population, the prevalence of AD shows an exponential increasing trend, intensifying the social, economic, and emotional impact on patients, their families, and healthcare systems (World Health Organization, 2023). In this context, early diagnosis and accurate classification of the disease stages become crucial, as they enable the implementation of therapeutic strategies that can delay symptom progression and optimize patient care planning (Alzheimer’s Association, 2022).

Pathophysiologically, Alzheimer’s disease is characterized by two primary neuropathological markers: the extracellular accumulation of plaques formed by beta-amyloid peptide (Aβ) and the formation of intracellular neurofibrillary tangles, resulting from the hyperphosphorylation of Tau protein (BERTRAM et al., 2005; HAINFELLNER et al., 1998). The aggregation of hyperphosphorylated Tau protein compromises cellular integrity and axonal transport, culminating in synaptic dysfunction and neuronal death (GUPTA et al., 2010). This dual cellular aggression process leads to progressive cerebral atrophy, which is macroscopically visible in Magnetic Resonance Imaging (MRI) and serves as one of the main biomarkers for the diagnosis and staging of the disease, justifying the use of computational methods for its detection.

Current diagnostic methods often integrate MRI images with the patient’s clinical history and neuropsychological tests. However, the interpretation of these images depends on the radiologist’s and physician’s experience, which can be particularly challenging in the early stages of the disease, where brain changes are subtle (Jack Jr. et al., 2018). Clinicopathological studies indicate that clinicians’ diagnostic sensitivity varies between 70.9% and 87.3%, and specificity between 44.3% and 70.8% (BEACH et al., 2012), highlighting the need for complementary tools that improve diagnostic accuracy and objectivity.

Artificial intelligence (AI), and more specifically computer vision, has emerged as a promising ally in diagnostic medicine. With the development of deep learning models, such as convolutional neural networks (CNNs), it has become feasible to create systems capable of identifying complex patterns in medical images with a level of precision that can be comparable or superior to that of human experts (Litjens et al., 2017). The application of CNNs in the analysis of MRI images for the detection and classification of AD has been documented in several works, demonstrating the potential of these algorithms to assist in diagnosis (Jo et al., 2019).

However, training a high-accuracy CNN from scratch requires a substantial volume of data, which is often a limitation in the medical context due to the scarcity of publicly available data. To circumvent this restriction, the methodology of transfer learning is recognized as a gold standard approach. This technique involves using a neural network pre-trained on a large dataset, such as ImageNet, and adapting it for a new specific task (SHIN et al., 2016). Architectures like ResNet, which incorporate residual connections to optimize gradient flow, have proven effective in this application, enabling the training of deep networks (HE et al., 2016; RAMALHO et al., 2024).

Given the growing prevalence of Alzheimer’s disease and the inherent limitations of traditional diagnostic methods, the application of advanced artificial intelligence techniques, such as transfer learning, presents a promising path for improving diagnosis. This work aimed to develop, train, and validate a deep learning model, based on the Transfer Learning methodology with the ResNet18 architecture, for the automatic classification of the four stages of Alzheimer’s disease: No Alzheimer’s, Very Mild Alzheimer’s, Mild Alzheimer’s, and Moderate Alzheimer’s, from MRI, aiming to create a highly accurate computational tool to support clinical diagnosis.

2. Material and Methods

This study was characterized as a research of an applied nature, with a quantitative methodological approach, focused on the development and validation of a deep learning model. The central objective was the automatic classification of the four stages of Alzheimer’s disease, namely: No Alzheimer’s, Very Mild Alzheimer’s, Mild Alzheimer’s, and Moderate Alzheimer’s, using brain magnetic resonance imaging (MRI) as the empirical object.

The dataset employed was obtained from the Kaggle data science platform, titled “Alzheimer’s Dataset (4 class of Images)” (SINGH, 2019). This public dataset comprised a total of 6,400 brain MRI images. The distribution of images per class was as follows: 3,200 images for No Alzheimer’s (50%), 2,240 for Very Mild Alzheimer’s (35%), 896 for Mild Alzheimer’s (14%), and 64 for Moderate Alzheimer’s (1%). The unbalanced distribution of classes was a factor considered in the data splitting and evaluation strategy.

For data preprocessing and structuring, the entirety of the 6,400 images was programmatically divided into three independent sets: training, validation, and testing. The `train_test_split` function from the Scikit-learn library (version 1.3.2) was used with stratification (70/10/20), ensuring the preservation of the original proportion of each class in all subsets. The sets comprised 4,480 images for training (70%), 640 for validation (10%), and 1,280 for testing (20%). To ensure reproducibility, a global random seed (seed = 42) was applied.

Additionally, data augmentation techniques were dynamically applied to the training set to mitigate overfitting and enhance the model’s generalization capability. The transformations included random rotations of up to 15 degrees and horizontal flips. All images were resized to 224×224 pixels. Subsequently, pixel values were converted to the [0, 1] range and normalized using the mean and standard deviation of the color channels from the ImageNet dataset, a standard procedure in transfer learning.

The development environment used was the Google Colaboratory (Colab) platform, Pro version, which provided an NVIDIA A100-SXM4-80GB Graphics Processing Unit (GPU). The development was carried out in the Python programming language (version 3.10.12), with the support of open-source libraries. The main libraries included PyTorch (version 2.3.0+cu121) for building and training the neural network, Torchvision (version 0.18.0+cu121) for accessing datasets and pre-trained architectures, Pandas (version 2.0.3) for tabular data manipulation, Matplotlib (version 3.7.1) and Seaborn (version 0.13.1) for graphical visualizations, and Numpy (version 1.25.2) for numerical computation.

The project’s core methodology was transfer learning, a consolidated strategy for scenarios with limited datasets (SHIN et al., 2016). A model pre-trained on a classification task of thousands of distinct objects from the ImageNet dataset was reused as a starting point. This approach allowed leveraging the network’s ability to extract generic visual features, accelerating learning for the new task and combating overfitting on a limited dataset.

The selected base architecture was ResNet18, a convolutional neural network with 18 layers (HE et al., 2016). The ResNet family, or Residual Networks, introduced residual connections (skip connections), which allow the output of a previous layer to be summed directly to the output of a subsequent layer. This mechanism creates a “shortcut” for information flow, optimizing gradient flow and enabling the training of very deep networks without the vanishing gradient problem.

The original architecture of ResNet18 was modified for the specific task. The original classification layer, designed for the 1000 ImageNet categories, was removed. In its place, a custom classification “head” was added, adapted for the four classes of Alzheimer’s disease. This “head” consisted of a dense layer with 512 input neurons and 128 output neurons, using the ReLU activation function. This was followed by a dropout layer with a probability of 0.20 (SRIVASTAVA et al., 2014) for regularization, and a final dense layer with 128 input neurons and 4 output neurons, corresponding to the number of classes, with a Softmax activation function.

The model training was performed in two distinct phases. In Phase 1, named Feature Extraction (10 epochs), the ResNet18 base was kept “frozen” (its pre-trained weights were not altered), and the training focused exclusively on the custom classification “head”. The Adam optimizer (KINGMA; BA, 2014) was used with an initial learning rate of 0.001. In Phase 2, Fine-Tuning (15 epochs), all model layers (ResNet base and custom head) were “unfrozen”, allowing adjustments to all weights with a drastically lower learning rate of 0.0001. This strategy aimed to refine the model for the specific task, avoiding “catastrophic forgetting” of pre-trained knowledge.

The loss function used to quantify the model’s error during training was Categorical Cross-Entropy (CrossEntropyLoss in PyTorch). The Adam optimizer (Adaptive Moment Estimation) was responsible for adjusting the model’s weights, adapting the learning rate for each network parameter, which contributed to faster and more stable model convergence.

The final model’s performance was rigorously evaluated on the test set, which consisted of data unseen by the model. The evaluation metrics calculated included overall accuracy, precision, sensitivity (recall), specificity, F1-Score, ROC curve, AUC (area under the curve), Youden’s index, Cohen’s Kappa coefficient, and Matthews correlation coefficient (MCC). The confusion matrix was also used to summarize the performance of the classification algorithm.

3. Results and Discussion

The evaluation of the deep learning model for the classification of Alzheimer’s disease stages, using brain magnetic resonance imaging (MRI), revealed exceptional performance, validating the transfer learning approach with the ResNet18 architecture. The development process was carried out in the Google Colaboratory environment, which, with the aid of an NVIDIA A100-SXM4-80GB GPU, allowed for a remarkably efficient training time of 2 minutes and 48 seconds. The final model incorporated a total of 11,178,564 trainable parameters, demonstrating the complexity and learning capacity of the adapted convolutional neural network.

The model training was structured in two distinct phases, aiming to optimize feature extraction and fine-tuning. In the first phase, which comprised the first ten epochs, the ResNet18 base was kept “frozen”, meaning its pre-trained weights were not altered. This initial stage focused on adjusting the custom classification “head”, allowing it to adapt stably to the generic features already learned by ResNet. Rapid learning was observed, with validation accuracy showing a consistent rise and loss progressively decreasing, indicating an effective adaptation of the classifier.

The second phase, initiated from the tenth epoch, involved the “unfreezing” of all model layers, including the ResNet base and the custom head, and the continuation of training with a drastically reduced learning rate (0.0001). This fine-tuning strategy resulted in a significant leap in performance, with the training and validation accuracy curves converging to levels above 95%. The proximity between these curves at the end of the process is a robust indicator that the model has reached an excellent generalization point, mitigating the risk of overfitting, a common challenge in deep learning models with limited datasets.

The analysis of the evolution of accuracy and loss during training confirmed the expected inverse correlation: as the loss decreased, the accuracy increased, in both the training and validation sets. The two-phase strategy for the learning rate, with a higher rate in the initial phase and a drastic reduction in the fine-tuning phase, was crucial for this smooth and effective convergence. Although the Mild Alzheimer’s class exhibited greater initial variability in F1-Score during fine-tuning, all classes demonstrated significant improvement and convergence to high performance, reflecting the robustness of the method.

The final model evaluation was conducted on a test set composed of 1,280 images, which were completely new to the model, ensuring an unbiased assessment of its predictive capability. The quantitative results obtained were exceptional, with an overall accuracy of 99.3%. This high value, along with a weighted average F1-Score of 0.993, confirms the model’s excellent balance across all classes, indicating that the system is capable of maintaining high precision and sensitivity simultaneously, even in a scenario of imbalanced classes.

The specificity, with a weighted average of 0.994, demonstrated the model’s high capacity to correctly identify negative cases, i.e., patients without Alzheimer’s, minimizing the occurrence of false positives. This aspect is particularly relevant in the clinical context, where an incorrect diagnosis can generate unnecessary anxiety and invasive procedures. The Youden’s index, which integrates sensitivity and specificity, also presented a high value of 0.987 in the weighted average, reinforcing the model’s excellent overall diagnostic capacity and its utility as a support tool.

Cohen’s Kappa coefficient, which corrects for chance agreement, reached 0.988, indicating an almost perfect agreement between the model’s predictions and the actual labels. This value is significantly high and above the threshold of 0.80, which already denotes an almost perfect agreement, attesting to the robustness of the model’s learning and that its results are not merely random or influenced by class imbalance. Complementarily, the Matthews Correlation Coefficient (MCC) of 0.988, considered one of the most informative metrics for imbalanced datasets, also pointed to an almost perfect prediction, taking into account all types of outcomes (true positives, true negatives, false positives, and false negatives).

The detailed analysis by class revealed that the “No Alzheimer” class obtained an accuracy of 0.988 and sensitivity of 1.000, with an F1-Score of 0.994. For “Very Mild Alzheimer’s”, the accuracy was 1.000 and sensitivity of 0.982, resulting in an F1-Score of 0.991. The “Mild Alzheimer’s” class presented an accuracy of 0.994 and sensitivity of 0.994, with an F1-Score of 0.994. A particularly notable finding was the perfect performance for the “Moderate Alzheimer’s” class, which achieved accuracy, sensitivity, specificity, F1-Score, Youden’s index, and AUC of 1.000 for all metrics, despite being the rarest class in the dataset, representing only 1% of the total images.

The AUC (Area Under the ROC Curve) values for all classes were extremely close to 1.0000, with “Mild Alzheimer’s” and “Moderate Alzheimer’s” reaching 1.0000, and “No Alzheimer’s” and “Very Mild Alzheimer’s” achieving 0.9999. These results indicate an almost perfect discrimination capability between classes, suggesting that the model is highly effective in distinguishing the different stages of the disease. The proximity of the ROC curves to the upper-left corner of the graph visually reinforces the model’s high sensitivity and specificity across all categories.

The visual analysis of the confusion matrix corroborated the numerical results, displaying a strong main diagonal and very low values outside of it. This demonstrates the model’s high accuracy across all classes. The few errors identified in the matrix confirmed a slightly greater difficulty in distinguishing between the “Mild Alzheimer’s” and “Very Mild Alzheimer’s” stages, which is an expected behavior and reflects the continuous nature of the disease’s progression. The perfect performance for the “Moderate Alzheimer’s” class, despite its minority representation in the dataset, is a particularly significant result, indicating the robustness of ResNet18 and the effectiveness of fine-tuning in learning distinctive features even with few samples.

Compared to the existing literature, the performance achieved by this model surpassed many reported results for Alzheimer’s disease classification using transfer learning on MRI datasets. Previous studies that applied approaches based on deep CNNs and transfer learning on datasets such as OASIS and ADNI, frequently reported accuracies in the range of 85% to 96% (ISLAM; ZHANG, 2020; SALEHI; ERDOGMUS; GHOLIPOUR, 2020). The accuracy of 99.3%, Kappa of 0.988, and MCC of 0.988 obtained in this study represent a substantial improvement.

The model’s superiority can be attributed to the strategic combination of the ResNet18 architecture, which is recognized for its efficiency and ability to handle the vanishing gradient problem through residual connections, and the two-phase fine-tuning methodology. This approach allowed the model to specialize in the task of MRI image classification for Alzheimer’s, leveraging knowledge pre-trained on a vast dataset and adapting it in a refined manner. The training stability and absence of significant overfitting, evidenced by the learning curves, are strong points that contributed to the final model’s high generalization capability.

The ability to perfectly classify the “Moderate Alzheimer’s” class, despite its extreme rarity in the training dataset, is a particularly relevant result. Many works in the literature face significant challenges with minority classes, where reduced representativeness can lead to training bias. The robustness of ResNet18, combined with the effectiveness of fine-tuning, allowed the model to learn distinctive features with few samples, demonstrating significant potential for early and accurate diagnosis, especially in more advanced stages, where identification is crucial for patient management.

In summary, the results obtained demonstrate that the developed deep learning model is highly effective in classifying the four stages of Alzheimer’s disease from MRI images. The high accuracy, Cohen’s Kappa, and MCC, along with the excellent discrimination capability between classes, validate the employed methodology and position the model as a promising tool to assist clinical diagnosis, contributing to early detection and patient care planning.

4. Conclusion

The present study aimed at the development, training, and validation of a deep learning model, employing the Transfer Learning methodology with the ResNet18 architecture, for the automatic classification of the four stages of Alzheimer’s disease from brain magnetic resonance imaging (MRI) scans. Exceptional model performance was observed, achieving an overall accuracy of 99.3% and a Cohen’s Kappa coefficient of 0.988 on the test set. The two-phase training strategy, involving feature extraction followed by fine-tuning, proved to be highly effective, enabling the model to learn complex patterns and generalize robustly. Notably, perfect classification capability was identified for the Moderate Alzheimer’s class, despite it being the rarest in the dataset, with a precision and sensitivity of 1.000. These findings position the model as a high-accuracy computational tool with great potential to support the clinical diagnosis of Alzheimer’s disease, contributing to early detection and patient care planning.

However, the study presents some limitations that should be considered. The model’s evaluation was performed on a single public dataset, suggesting the need for validation on external datasets from different imaging centers and acquisition protocols to confirm its real-world generalization. Additionally, the analysis was conducted on 2D images, while magnetic resonance imaging is a 3D modality, which may imply a loss of contextual information. For future studies, it is recommended to build models capable of directly processing 3D volumes and integrating multimodal data, such as computed tomography and blood biomarkers, aiming for an even more comprehensive and precise diagnosis. It is essential to emphasize that the developed model constitutes a diagnostic support tool, not a substitute for specialized clinical evaluation.

Bibliographic References

ALZHEIMER’S ASSOCIATION. 2022 Alzheimer’s disease facts and figures. Alzheimer’s & Dementia, 18(4): 700-789, 2022.

BEACH, Thomas G. et al. Accuracy of the clinical diagnosis of Alzheimer disease at National Institute on Aging Alzheimer Disease Centers, 2005–2010. Journal of neuropathology and experimental neurology, v. 71, n. 4, p. 266-273, 2012.

BERTRAM, L. et al. The genetic epidemiology of neurodegenerative disease. The Journal of clinical investigation, v. 115, n. 6, p. 1449-1457, 2005.

GUPTA, S. et al. Alzheimer’s disease: Genes, proteins, and therapy. Physiological Reviews, v. 81, n. 2, p. 741-766, 2001.

HAINFELLNER, J. A. et al. Coexistence of Alzheimer-type neuropathology in Creutzfeldt-Jakob disease. Acta neuropathologica, v. 96, n. 2, p. 116-122, 1998.

HE, K.; ZHANG, X.; REN, S.; SUN, J. Deep Residual Learning for Image Recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), p. 770-778, 2016.

ISLAM, J.; ZHANG, Y. Brain MRI analysis for Alzheimer’s disease diagnosis using an ensemble system of deep convolutional neural networks. Brain Informatics, v. 7, n. 1, art. 2, 2020.

JACK JR., C. R.; BENNETT, D. A.; BLENNOW, K.; CARRILO, M. C.; FELDMAN, H. H.; FRISONI, G. B.; HAMPEL, H.; JEGATHEESAN, A.; PETERSEN, R. C.; SNYDER, P. J.; SPERLING, R. A.; TEPPA, F. A NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimer’s & Dementia, 14(4): 535-562, 2018.

JO, T.; NHO, K.; SAYKIN, A. J. Deep Learning in Alzheimer’s Disease: Diagnostic Classification and Prognostic Prediction Using Neuroimaging Data. Frontiers in Aging Neuroscience, 11(220): 1-18, 2019.

KINGMA, D. P.; BA, J. Adam: A Method for Stochastic Optimization. In: Proceedings of the 3rd International Conference on Learning Representations (ICLR), 2015.

LITJENS, G.; KOOI, T.; BEJINARIU, B. E.; SETIO, A. A. A.; CIOMPI, F.; GHAFOORIAN, M.; VAN DER LAAR, J. A. W. M.; VAN GINNEKEN, B.; SÁNCHEZ, C. I. A survey on deep learning in medical image analysis. Medical Image Analysis, 42(1): 60-88, 2017.

ORGANIZAÇÃO MUNDIAL DA SAÚDE (OMS). Global status report on the public health response to dementia. Genebra: World Health Organization, 2021.

RAMALHO, B. A. C. et al. The impact of the orientation of MRI slices on the accuracy of Alzheimer’s disease classification using convolutional neural networks (CNNs). Journal of Medical Artificial Intelligence, v. 7, art. 35, 2024.

SALEHI, A. W.; ERDOGMUS, D.; GHOLIPOUR, A. Auto-Transfer Learning for Alzheimer’s Disease Diagnosis Using Structural MRI. In: International Workshop on Machine Learning in Medical Imaging (MLMI 2020), Springer, Cham, p. 193-201, 2020.

SHIN, H. C. et al. Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning. IEEE Transactions on Medical Imaging, v. 35, n. 5, p. 1285-1298, 2016.

SINGH, P. P. Alzheimer’s Dataset (4 class of Images). Kaggle, 2019. Disponível em: https://www.kaggle.com/datasets/preetpalsingh25/alzheimers-dataset-4-class-of-images. Acesso em: 14 out. 2025.

SRIVASTAVA, N. et al. Dropout: A Simple Way to Prevent Neural Networks from Overfitting. Journal of Machine Learning Research, v. 15, n. 1, p. 1929-1958, 2014.

Article originating from the Final Course Project of the Specialization in Data Science and Analytics of the MBA USP/Esalq

To learn more about the course, click here and access the MBX Academy platform

Most recent

Economy, Investments And Banking

October 23, 2026

Financial Education, Behaviors and Biases in Financial Decision Making

October 22, 2026

Challenges and Possibilities of the First School Year of Children with ASD for Teaching Work

October 22, 2026

Consumer Journey in the Decision to Purchase Infoproducts in the Digital Environment

October 22, 2026

Psychosocial Risks and Mental Health at Work in a Banking Institution

October 22, 2026

Digital Observatory of the Amazon: Data Science Applied to the Analysis of Social, Economic, Environmental, and Health Indicators

October 22, 2026

Continuous Learning as a Factor of Organizational Growth in the Perception of Collaborators

October 22, 2026

Statistical Regression in Microgrids for Operational Performance Estimation: Campusgrid Case

October 22, 2026

Forecast of Aggregate Default in the Brazilian Financial System by Sarimax Model with Macroeconomic Variables

October 22, 2026

Stakeholder Engagement Factors in International Technology Transfer to a Public Hospital

Personal financial management represents a growing challenge, influenced by cognitive biases that affect decision-making, even in individuals with high knowledge. The present study analyzed the relationship between the level of financial education and the occurrence of behavioral biases in the financial decision-making of adults residing in Brazil. An applied, quantitative, and descriptive-correlational field survey was conducted by administering a structured questionnaire to 163 respondents. The instrument included sociodemographic variables, financial knowledge and habits, and hypothetical situations to identify cognitive biases. The results indicated that higher levels of financial education were associated with better financial control, saving, and investment practices, with significant correlations with income and schooling. However, the presence of behavioral biases (anchoring, loss aversion, scarcity, optimism/negativity, present bias, status quo, overconfidence, and mental accounting) persisted even among individuals with higher schooling and income, with no statistically significant association with the analyzed sociodemographic variables. It was concluded that financial education, although essential for financial organization and control, was not sufficient to neutralize the influence of cognitive biases. The incorporation of behavioral approaches into financial education programs is suggested.

Keywords: Financial behavior; Behavioral economics; Personal finance; Cognitive heuristics; Cognitive biases.

October 22, 2026

Challenges and Possibilities of the First School Year of Children with ASD for Teaching Work

The inclusion of children with Autism Spectrum Disorder (ASD) in Early Childhood Education presents significant challenges, given the difficulty of adapting to new environments and the intensity of stimuli. The objective was to analyze the teaching performance in Early Childhood Education in the initial schooling process of children with ASD. A case study was conducted, using narrative interview as a data production technique. The results revealed that the teaching performance was characterized by mediation and affection practices, with the flexibility of routine and the use of playful resources as essential strategies for the child’s development and sensory organization. The importance of the school support professional for the success of the pedagogical work was also highlighted. The child’s integral development was promoted through intentional teaching actions that prioritized environmental adaptation, strengthening of bonds, and attention to the specific needs of ASD. As a contribution, a guide of pedagogical strategies aimed at welcoming children with ASD in public daycare centers was developed.

Keywords: Early Childhood Education; School Inclusion; Teacher Narratives; Inclusive Pedagogical Practices; Autism Spectrum Disorder.

October 22, 2026

Consumer Journey in the Decision to Purchase Infoproducts in the Digital Environment

The expansion of e-commerce and technological transformation have driven the need to understand consumer behavior regarding infoproducts, whose perceived experience has become crucial for professional development. The study aimed to analyze the factors that influenced the purchase decision of infoproducts in the digital environment, considering the stages of the consumer journey. An applied, quantitative, and descriptive research was conducted through a field survey with 100 consumers from Serra Gaúcha. Data collection occurred via a structured questionnaire, using non-probabilistic sampling. The results revealed a qualified audience with strong “mobile-first” behavior. It was found that the availability of detailed information and social approval were the main attraction triggers in the pre-purchase phase. In the transactional and subsequent stages, the minimization of operational friction, the agility of immediate access, and system security proved crucial for conversion and satisfaction, although gaps in human support efficiency were identified. Retention was based on the trust generated by positive past experiences, overcoming the price factor. It was concluded that the infoproduct consumer journey operates as an integrated system, requiring alignment between technological infrastructure, content curation, and support humanization to ensure engagement.

Keywords: E-commerce; Consumer behavior; Consumer experience; Loyalty; Virtual business.

October 22, 2026

Psychosocial Risks and Mental Health at Work in a Banking Institution

Mental health in the workplace is a topic of growing relevance, impacting professional development and employee well-being. A study was conducted to identify aspects and situations that can influence absenteeism levels in a banking institution. The research involved 108 participants and used a structured questionnaire, based on the short version of COPSOQ II by Gonçalves (2021), which contained 45 closed Likert scale questions and two open-ended questions. Data were collected via an online platform and analyzed with a descriptive and quantitative approach. The results revealed a scenario that indicated the need for adjustments in the distribution of activities and in managers’ perception of behaviors and signs of disorders. Important evidence was identified in the participants’ open-ended responses. It was concluded that the study contributed to promoting the perception of a safe environment for mental health in the organization.

Keywords: Absences due to disorders; Professional development; Operational leadership; Medical leave; Mental health.

October 22, 2026

Digital Observatory of the Amazon: Data Science Applied to the Analysis of Social, Economic, Environmental, and Health Indicators

The state of Amazonas faced historical challenges in consolidating the Unified Health System (SUS), stemming from geographical barriers, socioeconomic inequalities, and structural limitations. This study described the development and implementation of a health information geoportal, designed to support monitoring, evaluation, and planning in the Amazonian context. A descriptive case study approach was adopted, based on documentary analysis and secondary data, covering the mapping of functionalities, technical infrastructure, and validation steps of the tool. The Portal was developed with open technologies and integrated multiple national databases, offering interactive dashboards, thematic maps, and a digital document repository. Among the results, highlighted were the georeferenced visualization of key indicators, the possibility of temporal comparative analyses, and the use of the platform in reviewing health plans and redirecting strategic actions. The portal can represent an advance in digital governance and evidence-based management, with potential for replication in other highly complex territories, provided it is accompanied by investments in technological infrastructure, professional training, and inter-institutional coordination.

Keywords: Amazon; Health Evaluation; Digital Health; Health Information Systems.

October 22, 2026

Continuous Learning as a Factor of Organizational Growth in the Perception of Collaborators

The impact of continuous learning and knowledge management on organizational growth was studied in a context of globalization and constant changes. The study’s objective was to analyze employees’ perception of learning culture and its importance for personal and professional development. Data were obtained through an anonymous survey, with a quantitative approach, using a form with “survey” type questions applied to 87 professionals from the technology, health, education, commerce, and services sectors. The results indicated that employees had a positive perception of the contribution of continuous learning to personal and organizational growth, recognizing its benefits for the development of new skills, innovation, and competitiveness. However, it was observed that the incentive for internal knowledge exchange and the offering of practical training still presented moderate levels of agreement. It was concluded that continuous learning significantly contributes to individual and organizational development, suggesting that companies should develop strategic policies for learning focused on new skills, aiming for greater competitive advantage and innovation.

Keywords: Continuous learning; Self-development; Organizational growth; Knowledge Management; Innovation.

October 22, 2026

Statistical Regression in Microgrids for Operational Performance Estimation: Campusgrid Case

A multiple statistical regression model was proposed to estimate the operational performance of UNICAMP’s CampusGrid microgrid, based on internal electrical quantities. The microgrid integrates photovoltaic generation, batteries, and a gas generator. Energy management system data were collected at five-minute intervals over one month. The Operational Reliability Index, defined as the ratio between supplied and demanded energy, was established as the dependent variable. The model was adjusted for the islanded operation mode, evaluating the impact of demand, state of charge, solar generation, and supplied energy. The results revealed high explanatory power, indicating that operational reliability depends on the interaction between supply and demand. It was observed that the nighttime period concentrated the highest rate of partial supply, attributed to the reduction in solar generation. It was concluded that multiple regression offers an interpretable and useful framework for microgrid monitoring and control strategy support.

Keywords: Battery storage; Photovoltaic generation; Reliability Index; Islanded operation; Energy Management System.

October 22, 2026

Forecast of Aggregate Default in the Brazilian Financial System by Sarimax Model with Macroeconomic Variables

Aggregate non-payment in the Brazilian financial system is a crucial indicator of credit risk, the prediction of which is regulated by CMN Resolutions n.° 4.557/2017 and 4.966/2021 for the prospective measurement of expected losses. Aggregate non-payment was modeled and predicted using the SARIMAX model, projecting its trajectory until December 2026 under optimistic, base, and pessimistic scenarios. 164 monthly observations (July/2012 to February/2026) and macroeconomic variables such as unemployment, Selic, IPCA, IBC-Br, and household debt were used. The specification with contemporaneous variables was selected based on information criteria and coefficient stability, and residual diagnostics confirmed statistical adequacy, with a MAPE of 3.12%. Historical non-payment ranged between 4.33% and 7.64%. The estimation identified unemployment (β = 0.2534; p < 0.001) and Selic (β = 0.2128; p < 0.001) as the main positive determinants, with strong temporal persistence (AR(1) = 0.8965). Selic showed increasing correlation with non-payment (r = 0.62 at lag 12), while unemployment demonstrated a contemporaneous effect (r = 0.33). Projections indicated non-payment between 6.30% and 7.98% in December 2026, with a range of 1.68 percentage points, highlighting direct implications for provisioning in financial institutions.

Keywords: Unemployment; Predictive modeling; Monetary policy; Loss provisioning; Time series.

October 22, 2026

Stakeholder Engagement Factors in International Technology Transfer to a Public Hospital

This study described the critical factors perceived for stakeholder engagement in international technology transfer in a public hospital, with a focus on planning. The objective was to describe these conditioning factors, according to the perception of professionals. A single case study was conducted, applying a questionnaire to 25 professionals and using content analysis, complemented by descriptive tabulations. The main obstacles identified were resistance to change (72%), funding (64%), lack of communication (60%), and bureaucracy (40%). Only 28% of respondents recognized the existence of adequate tools for stakeholder management, while 48% denied it and 24% did not know. Exploratory cross-tabulations indicated variations in the assessment of influence according to time of practice, but did not establish causal relationships. Lack of engagement was associated with delays, low adherence, waste of resources, communication failures, compromised knowledge absorption, and risks to patient safety. The conclusions were limited to the perceptions of the researched group and supported preliminary recommendations for intersectoral communication and coordination.

Keywords: Case study; Project Management; Health innovation; Stakeholders; Public sector.

Economy

October 19, 2026

End of the 6×1 scale: what do the numbers allow us to conclude?

Following the promulgation of the end of the 6×1 shift system, the debate on the shorter workday moves to implementation and requires distinguishing social benefits from economic costs. The end of the 6×1 shift system transitioned from legislative discussion to a constitutional change in Brazil. On October 16, 2026, the National Congress promulgated Constitutional Amendment No. 141/2026, originating from Proposal […]

Artificial Intelligence

October 16, 2026

Why study if AI exists?

When answers become easy, learning comes to mean more than just arriving at them. Why do I need to learn this if I can ask an artificial intelligence? This type of question has accompanied education for decades. Why learn an operation if a calculator exists? Why memorize information if it can be found […]

Tax Management

October 15, 2026

Pink tax and the gender of price

How different marketing treatments for men and women are incorporated by consumption taxation Imagine a shelf on which two razors of the same brand and intended for the same purpose are displayed side by side: one has dark packaging and promises strength, precision, and efficiency; the other comes in shades of pink or […]

Health

October 13, 2026

The missing link in health transformation

Why is innovation not enough to drive changes in organizations and health systems Digital transformation has become a strategic priority for health systems worldwide. Electronic health records, artificial intelligence, telemedicine, remote monitoring, and data analysis are now part of the strategic agenda of governments and health organizations. Paradoxically, […]

Marketing

October 09, 2026

Bitcoin or pets: what Artificial Intelligence would not see

A real-life consumer intelligence case shows what gets left out of the brief A few years ago, an American insurance company hired a specialized consumer intelligence firm to answer three questions: The insurer was also considering launching paid insurance in bitcoin. At the time, the consultancy’s workflow was as follows: the analyst created a query […]

School Management

October 06, 2026

The school is not a casino: when trading cards become currency in the playground

Why, for certain students, does school manage to offer so few experiences of achievement? The World Cup is over. Brazil left the competition, the school holidays came, the final was played and the football calendar moved on. It would be reasonable to imagine that, with the end of the tournament and the return to classes, […] would also begin

Energy Transition

September 30, 2026

Renewable energies: the Brazilian advantage

Brazil needs to invest in the capacity to transform natural resources into innovation, productivity, and development. Talking about energy is talking about development. Behind industry, transportation, services, cities, and increasingly artificial intelligence, there is an energy infrastructure that sustains the economy. Understanding a country’s energy matrix is understanding […]

Data Analysis

September 30, 2026

When the data changes, has reality changed?

An analysis of almost 20 million teaching positions-year between 2020 and 2025 shows how changes in population composition and record production can alter the interpretation of historical series Between 2020 and 2025, the number of formal teaching employment positions identified in RAIS (Annual Relation of Social Information), from the Ministry of Labor, […]

Artificial Intelligence

September 28, 2026

Artificial intelligence and the trends

Why do so many people join in on the fun with ChatGPT? I was never the trend person. I still am not. Some pass through my timeline, I see them, I find them interesting but not to the point of trying them out. This was the case, inclusive, with some recent waves of images made by artificial intelligence, like the 80s trend, the one that […]

Tax Management

September 01, 2026

Tax reform and estate planning in the countryside

Efficiency in managing taxes facilitates governance and allows for the organization of family succession. Rural wealth organization has become a strategic measure in light of the tax reform in Brazil, which came into effect this year. The planning of activities and rural wealth management, through tools such as the transition from an Individual to […]

Renata de Gaspari Valdejão Almeida

Artificial Intelligence

August 05, 2026

Inside the company and off the radar: the risks of Shadow AI

According to Microsoft, 75% of professionals already use artificial intelligence at work without company approval. The use of generative artificial intelligence is already part of the corporate routine. Tools like ChatGPT, Copilot, and other platforms help professionals produce texts, analyze data, create presentations, and accelerate tasks in practically all sectors. But this advance […]

Renata de Gaspari Valdejão Almeida

Strategy

July 02, 2026

The future as a competence

Academic interest reinforces the projection of scenarios as an essential tool for decision-making. While some companies are still trying to get the next step right, others are already working with multiple possible futures — and preparing for all of them. But after all, what does it mean to “think about the future”? Is it a trend, a strategy, or a new competence that […]

Renata de Gaspari Valdejão Almeida

Innovation

Technology

May 21, 2026

Agro SP Corridor: innovation, institutional articulation and future

Pecege commands the executive operation of the 400 km ecosystem activities that became known as the Brazilian agro Silicon Valley In 2025, in addition to taking over the management of the Piracicaba Technology Park, the Pecege Institute was also designated to command the executive operation of the State Agribusiness Innovation Corridor activities, the so-called […]

Education

April 27, 2026

How to turn a TCC into a business

New management of the Piracicaba Technology Park offers consulting to MBA graduates The new project director of the Piracicaba Technology Park (PTP), Alexandre Barreto, often says that one of the greatest rewards he has ever had in life was seeing his MBA USP/Esalq students structure companies from their final projects […]

Technology

March 20, 2026

The human repertoire in the age of AI

Critical thinking and creativity are irreplaceable in the work of a marketing professional Artificial Intelligence (AI) is capable of incredible feats, but it does not yet possess the judgment, empathy, intuition, and strategic creativity necessary to do the job of a good marketing professional. In this area, it is important to know how to use the tool to automate […]

Renata de Gaspari Valdejão Almeida

Agribusiness

Innovation

February 23, 2026

ESG as a growth strategy in a cooperative in RS

Santa Clara balances social and environmental responsibility with innovation Founded in 1911 by Italian immigrants in Carlos Barbosa, in the Serra Gaúcha region, Cooperativa Santa Clara is an example that it is possible to balance economic growth with social and environmental responsibility. From a small cheese production company, the organization has evolved into a […]

Renata de Gaspari Valdejão Almeida

Agribusiness

Innovation

January 22, 2026

Brazil free of Foot-and-Mouth Disease: what changes for Brazilian beef?

With a herd of over 200 million head of cattle, the country is currently the leader in exports. Brazil recently received, for the first time, the certification of country free of foot -and-mouth disease without vaccination, conferred by the World Organisation for Animal Health (WOAH). The seal strengthens the country’s position as a reliable supplier of meat for the […]

Renata de Gaspari Valdejão Almeida

Innovation

December 09, 2025

Innovation: the new cycle of the Piracicaba Technology Park

Instituto Pecege assumes management, betting on the creation of new businesses Starting in 2026, Instituto Pecege will assume the management of the Piracicaba Technology Park (PTP). Among its responsibilities will be the administration, operation, maintenance, and expansion of the park, in addition to fostering activities in innovation, entrepreneurship, and science and technology. The choice was made […]

Renata de Gaspari Valdejão Almeida

July 06, 2026

Large-Scale Architecture: DSR-SCRUM Change Management

Mônica Alessandra Guerios; Maria do Carmo Assis Todorov Abstract generated by ResumeAI, an artificial intelligence solution developed by Instituto Pecege focused on synthesis and writing. Architecture and urbanism are configured as essential bases in an increasingly complex and challenging urban space, where the intrinsic connection between social, environmental, and technological issues […]