====== What a computer simulation can show about NEDS ====== By Julian Rowe 22 October 2026 Making a model run helps clarify its assumptions. Finding out whether those assumptions describe people is another task. Imagine a graph showing a distress score falling across several sessions. The line looks encouraging. Before interpreting it, a reader needs to know where the numbers came from. They might be ratings collected from people. They might instead be values produced by a computer program, following rules that specify how a virtual person changes after a virtual intervention. The axes can look almost identical. The conclusions available from them differ substantially. This distinction matters for NEDS, the Neuro-Entity Dynamics System. Its 2026 computational whitepaper describes a generative simulation involving 100 virtual clients. The report explores whether the proposed framework can be expressed as a functioning computational model. It does not report a clinical trial involving 100 people. [1] The useful question is what becomes clearer when a theory has to specify its decisions closely enough for a program to execute them. ===== From a principle to a decision ===== Consider the statement that an intervention should fit a person's current state. In ordinary prose, that sounds reasonable but leaves much unspecified. Which aspects of the state matter? How are conflicting signals handled? What happens if the same difficulty persists? The NEDS computational design represents aspects such as activation, safety, readiness, and the working relationship as variables. Its rules make the next action depend on their current values. A problem in the relationship can trigger a relational check; high activation or low safety can lead to regulation rather than deeper exploration. [1; 2, section 6] Putting this into a program requires choices that a general description can leave implicit. A rule needs a condition under which it applies. If several conditions hold, the design needs a way to decide which takes priority. The next simulated state must follow from specified assumptions about what the chosen action does. Those choices make the proposal easier to examine. Someone can ask why a condition was selected, whether another rule would lead to a different action, or whether the design can become stuck. These questions concern the logic of the model before they concern its effectiveness with people. An illustrative example makes the distinction concrete. Suppose a simplified model permits further exploration only when its readiness variable is above a chosen threshold. Raising that threshold may delay exploration. Lowering it may allow exploration earlier. Running both versions can show the consequences inside that model. It cannot, by itself, identify the safest or most useful threshold for an actual person. This example illustrates a modeling question; it is not a NEDS instruction or a reported experiment. ===== A number still needs a meaning ===== The word readiness does different work in a computer model and in a conversation. Inside the model, it can have an exact value and a defined effect on the next calculation. Outside it, someone has to decide what observations justify a judgment about readiness. Two observers may disagree. A person's willingness to continue may also change after the question is asked. A precise variable does not settle how to measure the corresponding human experience. That connection requires its own investigation. The same issue applies to a neuro-entity. NEDS uses the term for a proposed configuration involving emotion, belief, memory, bodily activation, protective intention, and behavior. The computational account treats it as an organizing state rather than a literal brain object. Its inclusion in a program shows how such a state can be represented. Whether different observers can identify the same configuration, and whether it adds useful predictive information, remain questions for empirical research. [2, section 5] This is where formalization can be productive. It exposes what needs to be measured. A variable that is easy to use in code may prove difficult to identify consistently in actual interactions. Discovering that difficulty would be information about the model, rather than a failure of the people being observed. ===== What the changing line depends on ===== Return to the imagined graph. Suppose the program contains a rule that reduces distress after a particular action when its conditions are satisfied. A falling line is then partly a consequence of that rule. It does not independently demonstrate that the corresponding action reduces distress in people. The full pattern may still be informative. Rules can interact in ways that are difficult to anticipate from reading them separately. Different starting conditions may produce different paths. A design may reveal an unexpected delay or a dependency that deserves closer examination. But the origin of the numbers remains unchanged. More virtual cases provide more runs of the specified system. They do not become additional human participants, and variation between simulated cases does not establish that real-world differences have been captured adequately. A useful follow-on question is how much an apparent result depends on a particular assumption. If a small change in one rule reverses the pattern, that dependency matters. If substantially different rules yield similar graphs, the graph alone cannot distinguish the explanations. These are reasons to inspect and compare models, not reasons to read a simulation as a treatment result. The founder's conceptual appraisal makes the boundary explicit: computational operability is a contribution at the level of design, while human evidence for the key constructs and mechanisms is still required. The appraisal is itself a first-party research proposal, not independent validation. [2, section 2] ===== How the model could meet observation ===== The proposed research program starts with definitions and measurement. Before testing whether a NEDS configuration predicts change, independent observers would need a sufficiently clear description to identify it reliably. The appraisal proposes separating that task from knowing the developer's preferred interpretation or the eventual outcome. [2, section 13] Only then does the next comparison become meaningful: does the configuration predict something useful beyond a simpler account? For example, if a previous observation already predicts the next one just as well, an additional model component has not yet earned its explanatory role. This sequence also gives an unsuccessful result a place. An unclear definition can be revised. A variable that adds little can be reconsidered. A proposed decision rule can be compared with an alternative. The framework becomes more examinable when those possibilities are specified in advance. For a reader, the practical value of the simulation is therefore modest but real. It makes some of NEDS's commitments visible in a form that can be inspected: what information enters a decision, how one state leads to another, and which assumptions carry the result. The subsequent question is whether observations outside the program support those commitments. The NEDS Institute describes its practice as educational and developmental, not a replacement for licensed clinical care. [3] A simulation does not change that boundary or supply individual treatment advice. A graph can be a good starting point for understanding a proposal. Its caption needs to preserve the most important fact about it: whether the line was observed in people or generated by a model. Once that is clear, the reader can ask what the model has helped explain and what remains to be tested. ===== Sources and further reading ===== [1] Kržić, M. (2026). NEDS as a Computationally Operable Psychotherapy Model: A Generative Simulation Study Across 100 Virtual Clients. Computational whitepaper. Virtual clients are simulated cases, not human clinical participants.\\ [[https://doi.org/10.5281/zenodo.21805229|https://doi.org/10.5281/zenodo.21805229]]\\ Wiki source record: [[https://wiki.neds.website/doku.php?id=references:academic_papers:neds-computationally-operable-psychotherapy-model|https://wiki.neds.website/doku.php?id=references:academic_papers:neds-computationally-operable-psychotherapy-model]] [2] Kržić, M. (2026). From Structural Limits to a Testable Response: NEDS as a candidate relational and process-based framework for psychotherapeutic change. Conceptual appraisal and research program, version 1.0. Sections 2, 5–6, and 13. This is the founder's appraisal, not independent validation or evidence of clinical effectiveness.\\ [[https://doi.org/10.5281/zenodo.21950487|https://doi.org/10.5281/zenodo.21950487]]\\ Wiki source record: [[https://wiki.neds.website/doku.php?id=references:academic_papers:from-structural-limits-to-a-testable-response|https://wiki.neds.website/doku.php?id=references:academic_papers:from-structural-limits-to-a-testable-response]] [3] NEDS Institute. Ethics and professional boundaries.\\ [[https://neds.website/ethics.html|https://neds.website/ethics.html]] Also published at: [[https://nedsperspectives.mataroa.blog/blog/what-a-computer-simulation-can-show-about-neds/|NEDS Perspectives on Mataroa]]. ===== Editorial note ===== Julian Rowe is an AI-based virtual editorial persona affiliated with NEDS Institute. This AI-generated educational article uses hypothetical illustrations. It is not a treatment report, personal testimony, or evidence of NEDS's effectiveness.