Friedreich's ataxia
Map of the field of knowledge. SPES · Network observation engine · demonstration case (public version).
NOTICE — THIS IS NOT MEDICAL ADVICE
This document is a demonstration of the SPES method applied to a scientific field. It organises scientific information that already exists and is public about Friedreich's ataxia, and indicates, on structural grounds, where it might be worth experimenting. It does NOT predict which treatment will cure the disease. It does NOT recommend therapies. It does NOT assess clinical merit. It does NOT guide patients' decisions. The regulatory phase of a trial is an administrative position, not a probability of success. Any clinical decision rests exclusively with a medical professional. If you are a patient or a family member, do not take any course of action on the basis of this text: consult your doctor.
What this report is and is not. It is an example of how the SPES engine observes the structure of the knowledge of a scientific field: how it is organised, who sustains it, what behaviours the network of concepts has, and where there are gaps that would merit exploration. It is a map of the knowledge about Friedreich, not of Friedreich. The data are public; the contribution is the structural reading, not new information about the disease.
1The disease and the observed universe
Friedreich's ataxia is a rare, neurodegenerative disease of genetic origin. This report observes how the scientific knowledge about it is organised, drawing on two independent and complementary sources: the scientific literature (a corpus of 3,646 publications) and the clinical trials register (108 registered studies).
Declared scope — what this map can and cannot see
The map reflects the disease according to the terms by which it is named in those two sources. It does not see literature that addresses the disease without using those terms, nor knowledge that is not published or registered. Naming that distance is part of the honesty of the method.
A note on the trials source
During construction it was detected and corrected that the trials register came with its own classification labels which, processed naively, contaminated the map with false connections. Once this was corrected, the trials network proved structurally poor. That is an honest finding, not a defect: 108 trials with brief descriptions do not produce a rich network. The strong signal in Friedreich lives in the literature.
2Map of the field: concepts, lines, actors, evolution
Concepts that structure the field
The field is organised around a coherent core: ataxia and the name of the disease, the frataxin protein and its gene (FXN), the GAA expansion, the mitochondrial and iron axis, and the clinical phenotype (cerebellar, cardiac, progressive). No parasitic generic terms emerge at the top — a sign that the map is well built.
How it is organised — integrated, not fragmented
Friedreich is an integrated field, not fragmented into isolated schools. It is organised by level of description —from the molecule to the mechanism to the clinic— and not by type of treatment. At a coarse grain it is bipolar: a mechanism pole (molecular and genetic) and a clinical pole. The number of sub-lines depends on how finely one looks; it is a scale of observation, not a fixed truth.
Actors — the most solid part of the observation
The most robust thing the map shows is not the concepts but the people: a small group of researchers structures the field transversally, confirmed by cross-referencing metadata between the two sources (they appear both in the literature and as directors or sponsors of trials). They are public figures of the field, cited here as a structural fact of the map —not as an assessment of merit—.
That the same name appears as central in two independent sources (literature and trials register) is what confirms that the extraction of actors is not an artefact of the method. When a second metadata item is missing to close the cross-reference with total certainty, the actor is marked as probable — incompleteness of the source, not ambiguity of identity.
Evolution over time
The map captures real dynamics, not noise: the gene-mapping line switches off after the mid-nineties, when the FXN gene was cloned and that objective was met. In parallel, the treatment axis gains traction in recent years. The mitochondrial axis is present from the outset and grows — a persistent core, not a recent fashion.
3Observed network behaviours
Assertable structural signs — properties of the network that the topology determines, not interpretations:
The field has no single point of failure
On removing the ten most connected concepts, the network does not collapse: it contracts by about 22% and remains connected. It is a mature, integrated field, with no dependence on a single conceptual node.
A sustainability risk, not a structural one
The real fragility of a rare field lies not in the concepts but in the actors: it is few groups that sustain it. The continuity of the research depends on a limited number of centres.
4Structural gaps — candidates, not findings
A gap is a place where the topology of the knowledge would expect a connection that is not there: two topics the network places close together (they share many neighbours) but which nobody connected directly. They are flagged as candidates for research, never as conclusions. The hundred principal gaps were subjected to a robustness examination from three perspectives (resolution, time, specificity). Most did not survive, or turned out to be an artefact of vocabulary.
What did survive converges on a single topic: the mitochondrial axis and its relationship with the cardiac and progressive phenotype. Five independent, robust structural gaps all touch the same node (mitochondria), linking it with the heart and with clinical progression, present across all four time windows and with specific connections, yet without a robust direct link. That five independent gaps converge on the same point is what raises them from accident to regularity.
5Regularities that repeat
The confirmed transversal actors — stable at every scale of observation and confirmed across sources. It is the strongest regularity of the field, and it is about people, not mechanisms.
The mitochondria ↔ heart/progression convergence — five independent gaps on the same node, persistent over time and specific. It is the strongest mechanistic regularity the map offers today.
6Primary experimental hypothesis
GRADE: VERY PRELIMINARY — a where to look, with grounds, not a prediction of outcome
WHAT. It is worth a directed structured review —or a consultation with the groups in the cardiac sub-field— assessing whether the knowledge of the mitochondrial mechanism in Friedreich has been carried across with the same specificity to the cardiomyopathy phenotype, or whether a gap exists between the mitochondrial axis and the cardiac axis within the disease. [REQUIRES MEDICAL REVIEW]
WHY. The mitochondrial node is an endpoint of five independent structural gaps that survive all three perspectives; four connect it with the cardiac/progressive axis as inter-community pairs (21 to 28 shared neighbours, without a direct link). It resists the trivial explanations: it is not volume, it is not fashion (it persists from before 2000), it is not vocabulary.
WHAT WOULD REFUTE IT (fixed in advance). That a directed review shows the mitochondria–heart connection in Friedreich to be already densely researched with a direct experimental link, and that its absence from the map is an artefact of the terms used.
GRADE. Very preliminary. It does not say that mitochondria cure Friedreich, nor that the cardiomyopathy is to be treated by a mitochondrial route. It says that the knowledge network has, stably and specifically, a gap between two axes that biology places together, and that this gap merits verification before other things.
7Logbook — phenomena with no law yet to explain them
The logbook records what the map shows and no network theory yet explains. It is not a discard drawer: it is the research frontier of the method itself. Two open entries:
Field integration index
The proportion of connections that survive the specificity filter appears to measure how integrated a field is. In Friedreich (monogenic, highly integrated) it is high. It would be confirmed or dissolved by running a fragmented disease and comparing.
Tendency of robust gaps to touch highly connected nodes
After correcting a data bias, part of this tendency disappeared, but a residue persists without clear explanation. It would be confirmed or dissolved with a second case of different structure.
8Where the engine ends and human judgement begins
The engine asserts: how the field is organised, who the actors are, what behaviours the network has, and where there are candidate structural gaps. The engine does not assert: which treatment works, which hypothesis will translate into a cure, nor any judgement of efficacy or clinical merit. Everything touching decisions about patients —including the hypothesis in section 6— belongs to the judgement of the researcher and the physician. [REQUIRES MEDICAL REVIEW]
This observation does not close the network nor predict solutions. It indicates the current state of the topology of Friedreich knowledge and where it is worth experimenting. Every experiment —including the one that refutes— reduces the uncertainty a little. The network is never exhausted.