AI EEG Analysis: Detection of Epileptic Discharges

We develop AI systems that automate EEG analysis, reducing review time by 4x and increasing sensitivity to rare patterns. Manual EEG analysis is a bottleneck in epilepsy diagnosis—a neurophysiologist spends 20-45 minutes on a 30-minute recording, and 24-hour monitoring is impossible to review completely without missing events. Our AI system flags suspicious epochs, reducing review time to 5-10 minutes. With 10+ years of experience in medical AI, we have delivered 50+ EEG analysis projects—from seizure detection to BCI.

How does AI detection of epileptic discharges work?

Detection of epileptiform activity involves searching for spikes, sharp waves, and spike-wave complexes. We use CNN+LSTM on EEG epoch time series. On public datasets CHB-MIT and TUH EEG, we achieve sensitivity 92-96% and specificity 86-91%. Our model outperforms classical machine learning methods (SVM, Random Forest) by 15-20% in F1 score. From our practice: a client project with 24-hour recordings—automatic annotation reduced neurologist time by 4x, uncovering 30% additional discharges missed during manual review. For a typical clinic, this translates to savings of $10,000 per month in neurologist time.

Why is EEGNet effective for EEG time series?

EEGNet is a compact convolutional network specifically designed for processing multichannel EEG. It uses depthwise and separable convolutions, greatly reducing the number of parameters. The model weighs 2-5 MB, almost 10 times less than transformers (e.g., LaBraM). This enables easy deployment on edge devices (ARM Cortex, Jetson). We adapt it to your electrode configuration (10-20 or 10-10 system).

# EEGNet — compact CNN specifically for EEG
class EEGNet(nn.Module):
    def __init__(self, n_classes, channels=64, samples=128):
        super().__init__()
        self.temporal_conv = nn.Conv2d(1, 8, (1, 64), padding=(0, 32), bias=False)
        self.bn1 = nn.BatchNorm2d(8)
        self.depthwise = nn.Conv2d(8, 16, (channels, 1), groups=8, bias=False)
        self.bn2 = nn.BatchNorm2d(16)
        self.separable = nn.Conv2d(16, 16, (1, 16), padding=(0, 8), bias=False)
        self.bn3 = nn.BatchNorm2d(16)
        self.dropout = nn.Dropout(0.5)
        self.fc = nn.Linear(16 * (samples//4), n_classes)

Comparison of architectures for EEG

What tasks do we automate?

Automatic Sleep Staging—labeling N1, N2, N3, REM, Wake by 30-second epochs. Cohen's Kappa 0.77-0.81, comparable to inter-rater agreement. Manual work takes 2-4 hours per night—with AI you get a ready hypnogram in minutes.

Anesthesia depth monitoring—we build custom ML models for propofol, isoflurane (different EEG patterns). Alternative to commercial BIS.

Brain-Computer Interface (BCI)—decode motor imagery and SSVEP to control prostheses, spellers, or exoskeletons.

Cognitive load and stress—neurofeedback, operator monitoring (aviation, nuclear plants): fatigue detection via EEG biomarkers.

Deliverables

Each project includes the following deliverables:

How we do it: stack and case study

We use PyTorch, Hugging Face Transformers, LangChain for pipelines. For time series—1D CNN/Transformer or time-frequency (STFT, wavelet) → 2D CNN. Spatial information is captured via electrode map (2D CNN or GNN).

From our practice: an ambulatory EEG monitor for seizure detection. EEGNet model (3.2 MB) runs on ARM Cortex-M4, latency <300 ms, battery lasts 48 hours. Sensitivity 94%, specificity 88%. Result: doctors receive a smartphone alert when a seizure is suspected. If your project requires a similar solution, contact us—we will prepare a prototype in 2 weeks.

Foundation Models for EEG—we use LaBraM, pre-trained on 25,000+ EEGs (TUEG). Fine-tuning on your data reduces the need for labeling to 10-20 hours. This is especially relevant if you have a limited volume of annotated recordings.

Developing an EEG analysis model: 5 steps

1. Data collection and preprocessing—load data, remove artifacts (ICA, ML classifiers), filter bands (0.5-70 Hz).
2. Architecture selection—compare EEGNet, CNN+LSTM, Transformer on latency/accuracy metrics.
3. Training and validation—k-fold cross-validation, hyperparameter tuning (learning rate, batch size).
4. Optimization—quantize to INT8, export to ONNX, test on target device.
5. Deployment and monitoring—integrate into your infrastructure, set up logging, A/B testing.

Timelines and cost

Timelines: from 4 to 16 weeks depending on complexity (task type, data volume, edge deployment requirements). Cost is calculated individually after analyzing your benchmark and requirements. We guarantee 12 months of support and free retraining when new data appears within the first year.

Have a project? Contact us for a consultation—our AI engineer will assess your data for free and propose the optimal solution.

Order a pilot project: a model prototype on your dataset in 2 weeks.

Industry AI Solutions: Healthcare, Finance, Retail, Manufacturing

We encounter the same pain points: a general text model doesn’t distinguish medical nomenclature, and a standard object detector confuses “weld seam scratch” with “casing scratch.” Each time these are different defects with different consequences. To avoid this, we build industry-specific solutions on top of general methods, but with deep domain knowledge — from regulatory requirements to data specifics. Over 5 years, we have completed 80+ projects in fintech, healthcare, retail, and manufacturing, and none were without adaptation to a specific business case.

Healthcare: Regulatory Maze and Data Governance

Medical AI differs not in technical algorithms but in a compliance-first approach. Depending on the country of application, the model may be a Class II or III medical device requiring clinical trials (FDA, CE MDR, GOST R). We ensure compliance with these standards at the architecture stage — fixing them post-factum is 10× more expensive.

Medical imaging. Detection on X‑rays, CT, MRI is a mature area. Models on ResNet, EfficientNet, SegFormer achieve AUC 0.94–0.97 on standard tasks (pneumonia on CXR, polyps on colonoscopy). Key issue is generalization: a model trained on data from one scanner manufacturer degrades on another due to differences in preprocessing and artifacts. Solution: domain adaptation via MONAI (Medical Open Network for AI) from NVIDIA, which includes DICOM loading, 3D augmentation, and confidence calibration. TotalSegmentator — for automatic segmentation of 117 structures on CT, production‑ready, Apache 2.0 license.

Clinical NLP. Extracting structured information from clinical records: diagnoses (ICD‑10/11), prescriptions, dates, indicators. medspaCy, scispaCy, MedCAT — specialized NLP libraries with ontologies (SNOMED‑CT, UMLS). Fine‑tuning BioBERT or ClinicalBERT on our data yields F1 0.85–0.92 on NER tasks versus F1 0.65–0.72 for general BERT. We verified this on a project with a regional oncology center — cancer stage extraction accuracy increased by 23%.

Clinical decision support. LLM assistants for clinical decision support are a regulatory gray area. We use an RAG system on top of clinical guidelines (UpToDate, local protocols) with explicit citation for each statement. The model does not diagnose but helps find relevant protocols. Stack: LlamaIndex + pgvector + pubmedbert-base-embeddings + Llama Guard for safety. Data in DICOM/HL7 FHIR, on‑premise deployment mandatory.

Finance: How to Ensure Interpretability of a Scoring Model under Basel IV?

The financial sector is one of the most mature in applying ML, but regulation is maximal. Every model affecting credit decisions falls under Basel IV, EU AI Act, GDPR Article 22. We deliver AI solutions for fintech that satisfy these requirements — in a project for a top‑10 bank we deployed a scoring model where each record required SHAP explanations.

Credit scoring. Gradient boosting (LightGBM, XGBoost) dominates. Neural networks yield +0.5–2% AUC but lose interpretability. Standard: LightGBM + SHAP to explain each decision. Fairness checking is mandatory: Fairlearn or aif360 for auditing disparate impact on protected attributes (age, gender). The default class is 1–5% — with an imbalance of 1:30, a model with 97% accuracy may have recall 0.2. Solution: focal loss, class_weight='balanced', SMOTE + careful validation. In one fintech scoring project, the model reduced credit losses by $2.1 million annually.

Algorithmic trading and risk management. LSTM and Transformer for price forecasting are popular but unstable in production due to non‑stationarity of financial series. A more robust approach: ML for signal generation (classification: up/down over horizon N) with traditional portfolio optimization on top. Backtesting via Zipline‑Reloaded, vectorbt, QuantLib. Proper backtesting is critical — look‑ahead bias kills results. We guarantee a clean experiment: all data at signal time is available in real time.

AML (Anti‑Money Laundering). Graph Neural Networks for analyzing transaction networks is an actively developing area. PyG, DGL for GNN. Task: detect suspicious patterns in transaction graphs (layering, structuring). Recall is more critical than precision — better 10 false alarms than miss one money laundering. In a project for a large payment service, we increased recall by 18% without increasing false positive rate.

Retail and e‑commerce: Recommendation Systems and Demand Forecasting

Recommendation systems. Current architectural standard: two‑tower model for retrieval + ranking with cross‑features. TensorFlow Recommenders or Merlin from NVIDIA for GPU‑accelerated feature processing. For small catalogs (<100k items), LightFM is sufficient. A common mistake is training on implicit feedback without accounting for position bias. Solution: IPW (Inverse Propensity Weighting) or randomized logging on a portion of traffic. Development time for a basic recommendation system is 4–8 weeks, including A/B test.

Demand forecasting and inventory optimization. Hierarchical forecasting: SKU → category → store → region. HierarchicalForecast from Nixtla automatically reconciles forecasts across levels. TFT or N‑HiTS for base forecast, gradient boosting for adjustment on exogenous factors (promotions, weather, events). One retail project led to a 15% reduction in stock‑outs due to precise promotion calibration.

Visual search and size compatibility. CLIP embeddings for image search — deploy in 2–3 weeks: clip‑ViT‑B‑32 or clip‑ViT‑L‑14, Faiss or Qdrant index, REST API. For size recommendation — specific models on return data and reviews with fit indication.

Manufacturing: Quality Inspection and Predictive Maintenance

Quality control and defect detection. CV models for product inspection are one of the most mature industry tasks. YOLOv10 for defect detection, SegFormer for segmentation. Specifics: class imbalance (defects are rare), high recall requirement (missing a defect is worse than false alarm). Typical dataset: 500–2000 defect images + 500–1000 normal. Few‑shot learning via DINO or SAM 2 works with 50–100 annotated examples. We gained experience on an electronics production line — recall 0.95 at FPR 0.03. A predictive maintenance deployment saved a manufacturing client $500,000 per year in unplanned downtime.

Predictive maintenance. Vibration sensors, current sensors, thermocouples → feature extraction → anomaly or mode classification. Models: LSTM‑AE for unsupervised, LightGBM for supervised (if failure history is available). Integration with SCADA/OPC‑UA via opcua-asyncio or MQTT. Key metric: False Negative Rate — a missed pre‑failure is more costly than a false alarm. Threshold tuned to business cost of each error type. Timeline: 3 to 6 months to production.

Digital twin and simulation. Surrogate models — ML models replacing expensive physical simulation. If a CFD simulation takes 6 hours and a surrogate (trained on 10,000 simulations) takes 0.01 seconds, that's 2,000,000× speedup for optimization. SALib for sensitivity analysis, botorch for Bayesian optimization on top of surrogate.

General Principles of Industry AI

Regardless of industry, there are patterns that work everywhere. Data matters more than architecture. In healthcare, 1000 quality labeled images are better than 100,000 poor ones. In manufacturing, 200 real defect examples are more valuable than 10,000 synthetic ones. Compliance‑first design — regulatory requirements are easier to embed into architecture from the start than to add later. Logging, explainability, versioning from day one. Domain expert on the team — an ML engineer without domain knowledge does slowly and error‑prone what an ML engineer plus a doctor/financier/technologist does quickly and correctly.

We guarantee certification to customer requirements (ISO 13485, SOC 2, GDPR) and provide full model documentation (model card, datasheet, compliance report). Our experience: 10,000+ engineering hours and 80+ projects.

Work Process for an Industry AI Solution

1. Domain immersion (2–3 days) — interviews with experts, studying regulatory requirements, auditing available data.
2. MVP design (1–2 weeks) — stack and architecture selection, feasibility assessment.
3. Development and validation (from 4 weeks to 6 months depending on industry) — model training, testing, compliance.
4. Integration and deployment (1–4 weeks) — on‑premise / cloud / edge, documentation, staff training.
5. Support and monitoring — model drift, retraining, SLA.

Estimated timelines:

Cost is calculated individually for each project. Get a consultation — we will evaluate your dataset, regulatory map, and business goals.

Why Choose Our Industry AI Solutions?

• 80+ completed projects in fintech, healthcare, retail, and manufacturing.
• 5 years on the market — proven experience with compliance and deployment.
• Quality guarantee: we ensure target metrics (AUC, recall, latency p99) and provide full documentation.
• Licensed technologies: PyTorch, MONAI, LightGBM, Qdrant — we use open‑source with commercially safe licenses.
• Flexibility: we work as a contractor or as an extension of your team.

Contact us for a free data audit and consultation. Request a proposal with a detailed work plan. We will discuss your task and prepare a commercial proposal.