Symptom Generation Mechanisms
The integrated model generates physiological state trajectories, but clinical relevance requires mapping these to symptoms. Symptom generation functions translate state variables to symptom severity scores:
\[ \begin{aligned} \text{Fatigue}(t) &= w_1 \cdot (1 - [\text{ATP}] \\/ [\text{ATP}]_\text{healthy}) + w_2 \cdot \mathbf{C}_\text{pro}(t) + w_3 \cdot (1 - F \\/ F_\text{healthy}) \\ \text{Pain}(t) &= w_4 \cdot \mathbf{C}_\text{pro}(t) + w_5 \cdot \mu_1(t) + w_6 \cdot [\text{ROS}](t) \\ \text{Cognitive}(t) &= w_7 \cdot \mathbf{C}_\text{CNS}(t) + w_8 \cdot (1 - [\text{DA}] \\/ [\text{DA}]_\text{healthy}) + w_9 \cdot (1 - \text{CBF} \\/ \text{CBF}_\text{healthy}) \end{aligned} \tag{1}\]
where \(\mathbf{C}_\text{pro} = [\text{IL-6}] + [\text{TNF-}\alpha] + [\text{IL-1}\beta]\) is the aggregate pro-inflammatory cytokine level, CBF is cerebral blood flow (proportional to cardiac output and cerebrovascular regulation), and \(w_1, dots, w_9\) are symptom-weighting coefficients estimated from patient-reported outcome correlations. These mapping functions are deliberately linear and phenomenological: they do not claim to model the neurobiology of symptom perception but provide a pragmatic interface between the mechanistic model and clinical observations.
The linear symptom generation functions are a major simplification. Symptom perception involves central sensitization, psychological modulation, and threshold effects that are not captured. The weighting coefficients (\(w_i\)) are likely patient-specific and may change over time (e.g., central sensitization increases pain perception for a given inflammatory stimulus). These functions should be interpreted as first-order approximations that enable model–data comparison, not as models of symptom generation per se.
Does the coupled multi-system model produce endogenous oscillations (symptom cycling) without external perturbation? The positive feedback loops (energy–immune, neuroimmune) combined with delayed negative feedback (cortisol, anti-inflammatory cytokines) create the mathematical conditions for limit cycle oscillations. If such oscillations emerge, they would provide a mechanistic explanation for the relapse–remission pattern observed in many ME/CFS patients, independent of identifiable triggers. Numerical exploration of the full model is required to determine whether the ME/CFS parameter regime produces stable limit cycles, chaotic dynamics, or a stable fixed point with amplified response to perturbations.