Healthy Exercise Response Dynamics
Before modeling the pathological response to exertion in ME/CFS, it is essential to model the healthy response. This serves two purposes. First, it provides face validation of the model: the same equations that will be used to describe ME/CFS pathophysiology must, with healthy parameters (\(\alpha_{\text{CI}} = 1.0\), \(\beta = 1.0\), \(\gamma = 1.0\)), reproduce the well-characterized dynamics of normal exercise physiology—a literature far more quantitatively constrained than ME/CFS. If the model fails this baseline test, it cannot be trusted for disease modeling. Second, it establishes the precise branch point where healthy and pathological responses diverge, clarifying which parameter changes convert a physiological process into a disease mechanism.
1 Healthy Exertion–Recovery Cycle
With healthy parameters, the model predicts a three-phase response to a moderate exertion pulse \(\phi(t)\) (demand exertion). The key quantity is the metabolic headroom ratio:
\[ R_{\text{headroom}} = \frac{J_{\text{production,max}} - J_{\text{demand,peak}}}{J_{\text{production,max}}} \tag{1}\]
In healthy individuals performing moderate exercise, \(R_{\text{headroom}} \approx 0.3\)–$ 0.5$: the ETC has substantial reserve capacity above peak demand. This headroom determines the entire downstream response.
Phase 1: Exertion (minutes to hours). ATP demand rises but remains within production capacity (\(R_{\text{headroom}} > 0\)). The ATP balance (atp balance) shows a transient dip—typically \(<10%\) from baseline—that is rapidly compensated by increased ETC flux and glycolytic upregulation. Lactate rises modestly (lactate dynamics) due to the glycolytic contribution but remains below the lactate threshold for moderate exercise. ROS production increases transiently (ros production) because ETC flux increases, but the increase is proportional and remains within antioxidant buffering capacity: \([\text{ROS}](t) < \text{ROS}_{\text{threshold}}\) throughout, so the damage indicator function in pem delay evaluates to zero. No cumulative oxidative damage accumulates.
Phase 2: Immediate recovery (0–2 hours post-exertion). As demand returns to baseline, lactate is cleared via the Cori cycle (\(v_{\text{clear}}\) term in lactate dynamics) with a half-life of approximately 15–30 minutes (Brooks 2018). ATP returns to baseline within minutes. ROS returns to steady-state levels as ETC flux normalizes. The metabolic perturbation resolves completely within 1–2 hours, consistent with the rapid metabolic recovery observed in healthy exercise studies.
Phase 3: Adaptation (6–96 hours post-exertion). The transient metabolic stress, although fully resolved at the whole-cell level, has triggered signaling cascades that operate on slower timescales. AMPK, activated during the ATP dip, initiates PGC-1\(\alpha\)-mediated mitochondrial biogenesis (biogenesis). Because the healthy cell has adequate \(\text{NAD}^\text{+}\) (\(\gamma = 1.0\)), the SIRT1 pathway is fully functional, and the biogenesis rate \(J_{\text{biogenesis}}\) is high. Simultaneously, the brief ROS transient—below the damage threshold but above resting levels—acts as a hormetic signal, upregulating antioxidant enzyme expression (a well-characterized phenomenon in exercise physiology). The net result over the subsequent 48–96 hours is a supercompensation: healthy mitochondrial mass \(M_h\) overshoots its pre-exercise value.
2 Supercompensation: The Healthy Adaptation Response
Supercompensation—the phenomenon whereby the system recovers to a state better than before the perturbation—is the fundamental mechanism of exercise adaptation. In the model, it emerges naturally from the mitochondrial dynamics equations (mito dynamics and biogenesis) when run with healthy parameters.
After a single exertion bout, the net change in healthy mitochondrial mass over the adaptation window \(\tau_{\text{adapt}}\) is:
\[ \Delta M_h^{\text{net}} = \int_{t_{\text{ex}}}^{t_{\text{ex}} + \tau_{\text{adapt}}} (J_{\text{biogenesis}}(s) - k_{\text{fission}} \cdot M_h(s) \cdot r_{\text{damage}}([\text{ROS}](s))) d s \tag{2}\]
In healthy individuals, \(\Delta M_h^{\text{net}} > 0\) because: (1) \(J_{\text{biogenesis}}\) is elevated post-exercise (AMPK-driven, with adequate \(\text{NAD}^\text{+}\)); (2) the damage term \(r_{\text{damage}}([\text{ROS}])\) is low (ROS remained sub-threshold during exertion); and (3) mitophagy efficiently removes any damaged mitochondria (adequate ATP for autophagy, mitophagy). The supercompensation window peaks at approximately 48–96 hours post-exertion, consistent with the well-established training adaptation timescale (Hood 2009) (Coffey and Hawley 2017).
Training adaptation is then modeled as iterated supercompensation. Repeated exertion bouts spaced within the supercompensation window produce monotonically increasing \(M_h\) (and therefore \(J_{\text{production,max}}\)):
\[ M_h^((n+1)) = M_h^((n)) + \Delta M_h^{\text{net}}(n) \quad \text{when } t_{n+1} - t_n \in [\tau_{min}, \tau_{max}] \tag{3}\]
where \(\tau_{min} \approx 24\) hours (minimum recovery) and \(\tau_{max} \approx 96\) hours (supercompensation window closure). This is the formal basis of the principle that training requires both adequate stimulus and adequate recovery—a principle that the model will later show is catastrophically violated in ME/CFS.
3 Delayed-Onset Muscle Soreness (DOMS): Healthy Delayed Symptoms
Healthy individuals also experience delayed symptoms after exertion: DOMS peaks 24–72 hours after unaccustomed exercise (Clarkson and Hubal 2002). This phenomenon shares the delayed timescale of PEM but differs mechanistically in ways that illuminate why PEM is pathological.
DOMS is driven by mechanical micro-damage from eccentric contractions, not by systemic metabolic crisis. The local damage triggers an inflammatory repair cascade:
\[ C_{\text{local}}(t) = C_{\text{peak}} \cdot \frac{(t - t_{\text{ex}})^a \cdot e^(-b(t - t_{\text{ex}}))}{t_{\text{peak}}^a \cdot e^{-b \cdot t_{\text{peak}}}} \tag{4}\]
where \(C_{\text{local}}\) is local cytokine concentration, \(t_{\text{peak}} \approx 24\)–$ 48$ hours, and parameters \(a\), \(b\) shape the rise-and-fall profile. Neutrophils infiltrate within hours, followed by macrophage-mediated debris clearance and satellite cell activation peaking at 48–72 hours (Peake et al. 2017). The process resolves completely within 5–7 days.
Three features distinguish DOMS from PEM:
- Local, not systemic: DOMS inflammation is confined to the exercised muscle. The systemic cytokine elevation is minimal and self-limiting. In PEM, the inflammatory response is systemic (Immune System Models).
- Self-limiting, no positive feedback: The DOMS inflammatory cascade follows a programmed rise-and-fall (doms inflammation) without the ROS \(->\) ETC damage \(->\) more ROS positive feedback loop that sustains PEM. The repair machinery has full access to ATP and \(\text{NAD}^\text{+}\).
- Repeated bout effect: Subsequent identical exertion produces progressively less DOMS (Clarkson and Hubal 2002), indicating successful adaptation. In ME/CFS, repeated exertion produces equal or worse responses (the energy ratchet, Disease Progression Models).
The DOMS model here is deliberately simplified, capturing only the inflammatory repair timeline to contrast with PEM. A complete model of eccentric exercise-induced muscle damage would require mechanical tissue damage terms, satellite cell proliferation kinetics, and extracellular matrix remodeling—beyond the scope of this metabolic framework.
4 The Branch Point: Where Health Becomes Disease
The healthy and pathological exercise responses diverge at a single identifiable point: whether exertion-induced ROS exceeds the antioxidant buffering threshold \(\text{ROS}_{\text{threshold}}\). This threshold determines whether the indicator function in pem delay activates:
- Below threshold (\(R_{\text{headroom}} >> 0\), healthy): ROS transient is buffered \(->\) no cumulative damage \(->\) biogenesis dominates \(->\) supercompensation \(->\) adaptation
- Above threshold (\(R_{\text{headroom}} \leq 0\), ME/CFS): ROS exceeds buffering \(->\) cumulative ETC damage \(->\) further reduced capacity \(->\) impaired repair (ATP-dependent mitophagy fails) \(->\) net mitochondrial loss \(->\) anti-supercompensation
This branch point can be expressed as a critical headroom ratio \(R_{\text{crit}}\) below which the damage indicator activates:
\[ R_{\text{headroom}} > R_{\text{crit}} => \Delta M_h^{\text{net}} > 0 \quad \text{(adaptation)}, \quad R_{\text{headroom}} < R_{\text{crit}} => \Delta M_h^{\text{net}} < 0 \quad \text{(damage)} \tag{5}\]
This constitutes a first validation of the model: the same equations that describe ME/CFS pathophysiology, when parameterized with healthy values, reproduce the well-characterized temporal dynamics of normal exercise physiology—immediate metabolic recovery in minutes, DOMS peaking at 24–72 hours, and supercompensation at 48–96 hours. These timescales are not fitted; they emerge from the interaction of metabolic kinetics (fast), inflammatory cascades (intermediate), and biogenesis (slow) that the model captures through its multi-timescale ODE structure.
Furthermore, the branch-point analysis provides the formal mechanistic basis for the clinical observation that graded exercise therapy (GET) harms ME/CFS patients (Geraghty, Hann, and Kurtev 2019): the same exertion stimulus that produces \(\Delta M_h^{\text{net}} > 0\) in health (training effect) produces \(\Delta M_h^{\text{net}} < 0\) in ME/CFS (progressive damage). The intervention is not merely ineffective—it is mathematically predicted to worsen the disease, because it repeatedly applies a stimulus that pushes the system further from recovery rather than toward it.
The defining feature of ME/CFS exercise intolerance is not the immediate metabolic response to exertion (which is qualitatively similar to healthy fatigue) but the failure of the post-exertional adaptation process: \(\Delta M_h^{\text{net}} < 0\) after exertion instead of \(\Delta M_h^{\text{net}} > 0\). This hypothesis predicts that: (1) mitochondrial DNA copy number decreases in ME/CFS patients 72 hours after standardized exercise but increases in matched healthy controls; (2) PGC-1\(\alpha\) target gene expression fails to upregulate post-exercise in ME/CFS despite adequate AMPK activation (indicating a SIRT1/\(\text{NAD}^\text{+}\) bottleneck); (3) interventions that restore \(\text{NAD}^\text{+}\) pools (e.g., NR supplementation) should partially restore the supercompensation response before improving exercise tolerance directly; and (4) the repeated bout effect (DOMS attenuation with training) is absent or inverted in ME/CFS patients.
Certainty: 0.50. The exercise physiology framework is well-established, and the mitochondrial dynamics model is internally consistent. The predictions are specific and testable by combining two-day CPET with pre/post mtDNA quantification and PGC-1\(\alpha\) gene expression profiling. However, direct measurement of in vivo mitochondrial biogenesis rates post-exercise in ME/CFS has not been performed.