# Dynamic Attentional Agents in Focused Attention Meditation: Hierarchical Computational Modeling of Expert-Novice Differences

**P. C. Kavi** · **Daniel Ari Friedman** · **G. Patow** (2026) · *Communications in Computer and Information Science, vol 2857 · Springer*

[![DOI](https://img.shields.io/badge/DOI-10.1007%2F978--3--032--16955--6__11-blue)](https://doi.org/10.1007/978-3-032-16955-6_11)

---

<!-- Schema.org structured data for search engines -->
<!--
{"@context":"https://schema.org","@type":"ScholarlyArticle","headline":"Dynamic Attentional Agents in Focused Attention Meditation","abstract":"A three-level hierarchical framework models attentional dynamics in focused attention (FA) meditation using the Free Energy Principle and Neuronal Packet Hypothesis. Thoughtseeds — transient, agent-like Markov-blanket entities — minimize variational free energy via bidirectional coupling with DMN, VAN, DAN, and FPN. Simulations reproduce expert–novice differences: experts achieve 49% lower free energy during breath focus, suppressed DMN activity, and faster distraction recovery.","keywords":["Active Inference","focused attention meditation","thoughtseeds","Free Energy Principle","hierarchical modeling","expert-novice differences","DMN","predictive processing","contemplative neuroscience"],"author":[{"@type":"Person","name":"P. C. Kavi"},{"@type":"Person","name":"Daniel Ari Friedman","url":"https://danielarifriedman.com/"},{"@type":"Person","name":"G. Patow"}],"isPartOf":{"@type":"Book","name":"Active Inference. IWAI 2025","publisher":"Springer","volumeNumber":"2857"}}
-->

*In: Albarracin, M., et al. (Eds.), Active Inference — IWAI 2025. Communications in Computer and Information Science, vol 2857. Springer, Cham.*

> **Note:** No freely available PDF for this chapter. Access via [Springer](https://doi.org/10.1007/978-3-032-16955-6_11) or institutional library.

## Abstract

> We develop a three-level hierarchical framework to model the attentional dynamics of focused attention (FA) meditation, laying a foundation for advanced active inference (AIF) implementations. Grounded in the Free Energy Principle and Neuronal Packet Hypothesis, we conceptualize meditation as a predictive processing system where "thoughtseeds"—transient, agent-like entities forming Markov blankets—minimize variational free energy via bidirectional coupling with attentional networks (Default Mode Network [DMN], Ventral Attention Network [VAN], Dorsal Attention Network [DAN], Frontoparietal Network [FPN]). Thoughtseeds, emerging from superordinate neuronal ensembles, compete for Global Workspace access, modulated by meta-cognitive precision weighting to stabilize attention. This model advances the Thoughtseeds Framework toward a computational phenomenology of Vipassana, setting the stage for integrated AIF formalisms and implementation, on top of rules-based statistical learning. Simulations across ranges of biologically plausible parameters—precision weighting (0.5 vs. 0.4), complexity penalties (0.2 vs. 0.4), learning rates (0.02 vs. 0.01)—reproduce patterns associated with expert-novice differences. Experts with higher precision weighting and higher learning rate notably achieve 49% lower free energy during breath focus, suppressed DMN activity (0.18 vs. 0.31), and faster distraction recovery, consistent with neuroimaging findings. Bidirectional message passing enables bottom-up formation of attentional states and top-down constraint of dynamics, offering a mechanistic account of expertise as optimized precision allocation. This framework provides testable predictions for meditation skill development, with future extensions planned to enhance AIF rigor and computational phenomenology for applications in contemplative neuroscience, computational psychiatry, and cognitive training.

## Keywords

`Active Inference` · `focused attention meditation` · `thoughtseeds` · `Free Energy Principle` · `hierarchical modeling` · `expert-novice differences` · `DMN` · `predictive processing` · `Neuronal Packet Hypothesis` · `contemplative neuroscience`

## Key Contributions

- A three-level hierarchical framework for focused attention meditation grounded in the Free Energy Principle and Neuronal Packet Hypothesis.
- Formalization of thoughtseeds as transient, agent-like entities with Markov blankets competing for Global Workspace access via meta-cognitive precision weighting.
- Biologically plausible simulations reproducing expert–novice differences: 49% lower free energy during breath focus, suppressed DMN activity (0.18 vs. 0.31), and faster distraction recovery in experts.
- Testable predictions for meditation skill development; pathway to integrated AIF formalisms for contemplative neuroscience and computational psychiatry.

## Methods

- Three-level hierarchical model: superordinate neuronal ensembles → thoughtseed agents (Markov blankets) → attentional networks (DMN, VAN, DAN, FPN).
- Bidirectional message passing: bottom-up attentional state formation and top-down precision-weighted constraint.
- Simulation across biologically plausible parameter ranges: precision weighting, complexity penalties, learning rates.
- Comparison to neuroimaging findings on expert meditators.

## 🎯 Consulting & Tutoring

**Available for AI Research Consulting and Tutoring.** [Contact Daniel Ari Friedman, PhD](https://danielarifriedman.com/) for collaboration on Active Inference, contemplative neuroscience, and hierarchical modeling.

## Citation

```bibtex
@inproceedings{Kavi_2026_FocusedAttentionMeditation,
  author    = {P. C. Kavi and Daniel Ari Friedman and G. Patow},
  title     = {{Dynamic Attentional Agents in Focused Attention Meditation: Hierarchical Computational Modeling of Expert-Novice Differences}},
  booktitle = {Active Inference. IWAI 2025},
  series    = {Communications in Computer and Information Science},
  volume    = {2857},
  publisher = {Springer, Cham},
  editor    = {Albarracin, M. and others},
  year      = {2026},
  doi       = {10.1007/978-3-032-16955-6_11},
}
```

## File Inventory

- `AGENTS.md`
- `README.md`
- `SKILL.md`

> No PDF is freely available for this chapter. Access via [Springer](https://doi.org/10.1007/978-3-032-16955-6_11).