Targeting Compensatory Motor Circuits in PLS

A PLS patient who had not used cannabis for two years tried it a couple of times and experienced a dramatic acute worsening of walking that lasted for hours. Most online resources would file this under "cannabis has side effects, be careful." This page treats it as something more useful: a natural experiment, with real implications for which brain circuits are holding the patient's walking together — and therefore which compounds might reinforce them.

The clinical observation

The observation is simple but the implication is not. After roughly two years of abstinence, a PLS patient smoked high-THC cannabis on a couple of occasions and, within a short time on each occasion, his walking deteriorated dramatically. Not subtly — dramatically. The effect resolved over the following hours as the drug cleared his system, and the underlying disease was unchanged.

For someone who has spent time caring for or reading about PLS, the automatic response might be: of course THC worsens motor function, cannabinoids impair coordination, this is well-known. That response is technically correct but analytically lazy. It treats the acute worsening as a generic "cannabis side effect," when the specific pattern — the magnitude of impairment, the circuit it operated through, the disease context that made it visible — is actually informative.

The useful clinical observation is not "THC makes walking worse in PLS patients." The useful observation is: in this patient, who has substantial corticospinal tract damage, the acute disruption of CB1-rich motor circuits by THC was sufficient to make walking fail. That tells us something about how much of his walking depends on those circuits. And that in turn tells us something about what might happen if we targeted those same circuits with compounds that enhance rather than suppress them.

This page is a scientific reasoning exercise. It moves in one direction from the THC observation — backward through the pharmacology to the anatomy, and forward from the anatomy to the specific compounds that might be worth trying. The reasoning is mechanistic and honest about its limits. It does not claim certainty. It claims that the circuit logic is coherent enough to be worth a conversation with a neurologist.

What THC actually tells us about the brain

THC (delta-9-tetrahydrocannabinol) is a partial agonist at CB1 cannabinoid receptors. This places it in a class of drugs with an unusual anatomical specificity: unlike most neurotransmitter receptor families, CB1 is not expressed uniformly across the brain. It is densely concentrated in particular structures and nearly absent in others. The result is that THC does not produce a diffuse "slowing of the brain." It produces a pattern of effects that maps onto the specific regions where CB1 is highly expressed.

CB1 receptors are predominantly presynaptic. When THC activates them, the primary effect is reduced neurotransmitter release from the terminal button. Both glutamate and GABA release can be suppressed, depending on which presynaptic terminal carries the CB1 receptor. The net functional effect — whether a circuit becomes more or less active — therefore depends entirely on the circuit architecture. In circuits where CB1 sits on excitatory inputs, activation blunts excitation. In circuits where CB1 sits on inhibitory inputs, activation blunts inhibition and can paradoxically increase the downstream output. The endocannabinoid system is a modulator of gain, not a simple on/off switch, and the direction of gain change is circuit-specific.

For motor function, this matters because the main motor-relevant brain regions have very different CB1 densities and circuit architectures. THC does not act equally on all of them. It hits some regions hard, barely touches others, and the regions it hits hardest are not necessarily the ones that do the most work in a healthy person. In a healthy person, the corticospinal tract is the dominant motor pathway — it carries the primary voluntary movement signal from motor cortex to spinal motor neurons — and it is relatively CB1-poor along its length. This is why healthy people who smoke cannabis can usually still walk, despite obvious impairments in coordination, timing, and balance.

In a PLS patient, however, the corticospinal tract is the damaged structure. And the question this page is asking is whether the CB1-rich circuits — the ones THC hits hard — have become disproportionately important in sustaining whatever residual walking ability the patient has. If they have, then THC's acute suppression of those circuits would produce exactly the dramatic deterioration that was observed. It would not be a general drug effect. It would be a specific pharmacological probe revealing which circuits are doing the compensatory work.

Where CB1 receptors live (and why it matters)

The anatomy here is well-characterized. The cerebellum has the highest CB1 receptor density in the entire brain. The expression is not uniform across cerebellar cell types — it is concentrated at the parallel fiber synapses onto Purkinje cells in the molecular layer, and on the basket and stellate interneurons that regulate Purkinje cell firing. The Purkinje cell is the output neuron of the cerebellar cortex: its axon is the final common pathway through which all cerebellar computation — all the error correction, timing precision, and motor coordination that the cerebellum produces — travels before reaching the cerebellar nuclei and from there the motor thalamus, motor cortex, and brainstem. When THC activates CB1 receptors on parallel fiber terminals, it blunts the excitatory drive onto Purkinje cells. When it activates CB1 on interneuron terminals, it blunts inhibitory regulation of Purkinje firing. The net effect is that the Purkinje cell's characteristic precision — its role as the biological clock and error-correction engine of the motor system — is acutely degraded.

The basal ganglia also have high CB1 density, particularly in the substantia nigra pars reticulata and globus pallidus, both of which are GABAergic output structures. The striatum has moderate CB1 expression on corticostriatal glutamatergic inputs. This distribution means that THC can influence both the direct and indirect pathways through which the basal ganglia regulate motor output. The specific effect on movement involves reduced basal ganglia facilitation of desired movements and potentially increased suppression of competing movements, though the exact functional impact varies by dose and individual physiology. What is consistent is that basal ganglia motor gating — the circuit's contribution to initiating and sustaining movement — is impaired by acute THC exposure.

The motor cortex itself has moderate CB1 expression, contributing to the coordination and timing impairments that cannabis produces in healthy subjects. This effect is real but secondary relative to the cerebellar and basal ganglia contributions.

The dorsal horn of the spinal cord has moderate CB1 expression, but this is predominantly on pain-modulating afferent terminals — it is the circuit through which cannabinoids reduce pain and modulate spasticity, not through which they produce motor coordination impairment. The spasticity-reduction seen with nabiximols in the CANALS trial likely operates partly through this spinal mechanism.

And then there is the corticospinal tract itself. The axons running from motor cortex through the corona radiata, internal capsule, cerebral peduncles, and brainstem pyramids to the spinal cord express relatively low CB1 along their length. THC does not primarily impair corticospinal axon conduction. It disrupts the cerebellar, basal ganglia, and cortical circuits that generate, modulate, time, and gate the signal that the corticospinal tract carries. In a healthy person, where the corticospinal tract dominates voluntary movement, this circuit-level disruption is a moderate impairment overlaid on a robust primary pathway. In a PLS patient, the situation is inverted.

The motor reserve and compensatory circuit hypothesis

In a healthy person, normal walking uses perhaps a third of available motor capacity. The corticospinal tract produces the primary voluntary movement commands. The cerebellum continuously monitors movement against expectation, corrects errors in real time, and ensures the precise timing that distinguishes smooth gait from stumbling. The basal ganglia gate and initiate movement, suppress competing motor programs, and sustain the motor "intention" that keeps walking going once started. The brainstem reticulospinal system provides descending facilitation to spinal central pattern generators, contributing to the automatic, rhythmic aspects of walking that do not require conscious attention. All of these systems have surplus capacity relative to what normal walking demands.

In PLS, the corticospinal tract is the primary site of damage. Betz cell loss in layer V of primary motor cortex progressively reduces the size and reliability of the primary voluntary motor signal. What happens next is not that walking stops — it is that walking increasingly recruits the compensatory capacity of the systems that remain. Spared motor cortex regions adjacent to the lesioned areas can contribute corticospinal output. Cerebellar error-correction must work harder against a noisier primary command. Basal ganglia gating must compensate for reduced command clarity by more actively filtering movement initiation. The brainstem reticulospinal system, which is largely spared in PLS because it originates in structures not primarily targeted by the disease, takes on a greater share of descending motor drive.

The "motor reserve" — the margin between what the motor system can produce and what walking requires — shrinks as the disease progresses. But critically, what remains in that reserve is no longer the original corticospinal-dominant architecture. It is a shifted system, one that has disproportionately recruited exactly the CB1-rich compensatory circuits: cerebellar Purkinje cell output, basal ganglia motor loops, brainstem projections. The normal redundancy that allowed a healthy person to tolerate THC-induced disruption of these circuits without catastrophic motor failure is gone. The compensatory circuits are not backup systems operating alongside a healthy primary pathway. They are the primary pathway, or something close to it.

When the patient smoked high-THC cannabis, the drug activated CB1 receptors on parallel fiber terminals in the cerebellar molecular layer and on GABAergic terminals in the basal ganglia. The circuits that were carrying his walking acutely lost precision and gain. The motor reserve, already thin, went negative. Walking failed — not because the disease had progressed, but because the pharmacological probe had temporarily deactivated the compensation that was maintaining function.

This hypothesis has not been tested in a controlled study specific to this scenario. But it is grounded in a coherent reading of CB1 anatomy, PLS pathophysiology, and the pharmacology of motor compensation. The compounds that follow are selected on this basis. Not every PLS patient will respond, and not every PLS patient has the same degree of compensatory circuit recruitment. But for a patient who has already, inadvertently, demonstrated that their walking depends heavily on CB1-rich circuits, the hypothesis is unusually well-supported by direct clinical observation.

Compounds that target cerebellar Purkinje cells

The leading compound in this category is dalfampridine — known in Europe as Fampyra, in the United States as Ampyra, and in the pharmacological literature as 4-aminopyridine (4-AP). It is the compound with both the most mechanistically relevant action on cerebellar function and the most directly relevant clinical trial evidence in PLS specifically.

Dalfampridine's primary mechanism is blockade of voltage-gated potassium channels, particularly Kv1 channels. In the context for which it is approved — walking impairment in multiple sclerosis — the textbook explanation is that K+ channel blockade in demyelinated axons prolongs the action potential and partially restores conduction through segments where the myelin sheath has been lost. This is true, and it is the mechanism most neurologists know. But it is not the only mechanism, and for PLS it may not be the most important one.

The work of Kalla, Glasauer, and colleagues, published in Brain in 2007 (vol. 130, pp. 2441–2451), showed that 4-aminopyridine restores neural integrator function in downbeat nystagmus by directly enhancing Purkinje cell pacemaking precision. Downbeat nystagmus is an eye movement disorder caused by dysfunction of the cerebellar flocculus, whose Purkinje cells normally suppress unwanted downward eye drift. When those Purkinje cells fire irregularly or too slowly, gaze stability is lost. 4-AP restored the regularity of Purkinje cell firing — specifically by prolonging the action potential and restoring the precision of the afterhyperpolarization that paces the cell's intrinsic rhythmicity. This is a Purkinje cell effect, not an axon conduction effect. Kalla and Glasauer subsequently showed similar results in mouse models of spinocerebellar ataxia type 6 in a 2016 Scientific Reports paper, where 4-AP reversed the ataxic phenotype by restoring Purkinje cell firing patterns.

The clinical indications where dalfampridine has real evidence now include: episodic ataxia type 2 (EA2), where it is considered first-line ahead of acetazolamide in some centers; downbeat nystagmus; the recently characterized GAA-FGF14 disease (published in eBioMedicine in 2024, a previously unrecognized form of late-onset cerebellar ataxia that responds strikingly to 4-AP); and spinocerebellar ataxia type 6. All of these indications share the same underlying mechanism: dysfunctional Purkinje cell firing, restored by K+ channel blockade. The compound has found its niche in cerebellar disorders, and that niche is precisely mechanistically relevant here.

In multiple sclerosis, the Phase 3 ENHANCE trial demonstrated that roughly 40 percent of treated patients met the responder threshold on walking speed (Timed 25-Foot Walk), with responders showing approximately 25 percent improvement in walking velocity. Not every MS patient responds — but those who do, respond meaningfully. This responder-versus-non-responder pattern appears in cerebellar ataxia data as well: Purkinje cell dysfunction is not uniform across patients, and the magnitude of K+ channel dysfunction that 4-AP can correct likely varies between individuals.

The most directly relevant evidence for PLS is the Weill Cornell trial, registered as NCT02868567. This was an 18-week open-label trial of dalfampridine in 35 patients with primary lateral sclerosis or upper-motor-neuron-dominant ALS, using the Timed 25-Foot Walk as the primary endpoint. The trial reflects real clinical interest in exactly the application being discussed here — a PLS-specific investigation of the compound, in a population selected for upper motor neuron pathology, using a functional walking outcome. Full peer-reviewed results have not been widely published as of this writing, but the trial's existence and design are themselves a signal that knowledgeable clinicians at a leading MND center considered the mechanistic case compelling enough to study formally. The dedicated page for this trial covers what is publicly available from the registry and any subsequent publications.

Practically: Fampyra is approved in France by the ANSM for walking impairment in multiple sclerosis. Prescribing it off-label for PLS or upper-motor-neuron-dominant disease requires a neurologist willing to engage with the reasoning, but it is a realistic conversation at centers familiar with both MS and MND. The standard dosing is 10 mg twice daily (extended release), taken at least 12 hours apart. The main safety consideration is seizure risk, which is low at standard doses but increases in patients with pre-existing epilepsy or significantly impaired renal function.

The second cerebellar stabilizer in this category is acetazolamide. Where dalfampridine acts by prolonging the action potential through K+ channel blockade, acetazolamide — a carbonic anhydrase inhibitor — stabilizes Purkinje cell excitability through an entirely different route: by modulating intracellular pH and the associated shifts in potassium conductance that follow from it. These are different aspects of Purkinje cell physiology, which is why some patients respond to one but not the other, and why they are not simply interchangeable.

Acetazolamide is the first-line treatment for episodic ataxia type 2, where it produces dramatic reduction in the frequency and severity of attacks in most patients — this is one of the more reliable pharmacological responses in all of neurology. It has decades of safety data from its use in glaucoma, altitude sickness, and periodic paralysis. There are no PLS trials of acetazolamide, and the evidence for progressive (as opposed to episodic) cerebellar dysfunction is weaker than for dalfampridine. But the mechanism is coherent, the safety profile is well-established, and it represents a reasonable second-line option if dalfampridine does not produce a clear response. Available in France as Diamox. More detail at the dedicated study page.

Riluzole deserves a brief mention here, not because of its ALS indication but because small trials have investigated it as a cerebellar ataxia treatment, including spinocerebellar ataxia types 1, 2, 3, and 7, with modest positive signals on some cerebellar outcome measures. Riluzole's mechanism in this context is likely related to glutamate modulation in the cerebellum rather than sodium channel blockade. Given that many PLS patients are already on riluzole for its possible neuroprotective effects, it is worth noting — but its contribution to cerebellar compensation is likely too weak to be the primary treatment target here.

Compounds that target basal ganglia output

The leading compound for basal ganglia motor circuit enhancement is ropinirole, a dopamine D2 and D3 receptor agonist widely known for its use in Parkinson's disease and restless legs syndrome. In France it is available as Adartrel (for restless legs, at lower doses) and Requip (for Parkinson's). It is generic, inexpensive, and familiar to virtually every neurologist in the country.

The mechanism relevant here is dopaminergic enhancement of basal ganglia motor output. The direct pathway through the basal ganglia — striatum to globus pallidus interna to thalamus to motor cortex — is facilitated by dopamine acting at D1 receptors in the striatum. The indirect pathway, which suppresses competing motor programs, is modulated by dopamine at D2 receptors. Ropinirole acts primarily at D2 and D3 receptors, and in the motor domain its dominant effect is to improve the basal ganglia's ability to gate and initiate desired movements. This is the circuit that THC partially disrupts through CB1 activation on GABAergic terminals in the substantia nigra and globus pallidus. Ropinirole works through a different receptor system on overlapping circuits.

The most significant recent evidence for ropinirole in motor neuron disease comes from the ROPALS trial, conducted by Morimoto and colleagues at Keio University, published in Cell Stem Cell in June 2023. This was a Phase 1/2a, double-blind, placebo-controlled trial in 20 patients with sporadic ALS, randomized to ropinirole or placebo for 24 weeks followed by an open-label extension.

The headline result was a 27.9-week average delay in disease progression in the ropinirole group compared to placebo. Treated patients showed better mobility, muscle strength, and lung function, and better survival outcomes in the open-label extension period. These are clinically meaningful results for a disease with few effective therapies — they are not marginal.

What makes ROPALS scientifically distinctive is not just the efficacy signal but the biological rationale behind patient selection and the mechanism story. Morimoto's team generated induced pluripotent stem cells (iPSCs) from the blood of each enrolled patient, differentiated them into motor neurons, and studied the cellular phenotype of each patient's motor neurons before they were randomized. ALS motor neurons from these iPSC cultures showed abnormal morphology, dysregulated gene expression, and characteristic metabolite patterns. Ropinirole treatment of the iPSC-derived motor neurons in culture reversed multiple aspects of these abnormalities — not through a generic neuroprotective mechanism, but through something that appeared specific to the dopaminergic circuitry around which those motor neurons are embedded.

This is the part of the ROPALS story that elevates it beyond a symptomatic dopaminergic effect. The compound appears to have a cellular protection effect on motor neurons themselves, in addition to whatever symptomatic benefit comes from enhancing basal ganglia motor output. Two plausible mechanisms, one drug. The iPSC work also identified a metabolomic signature that predicted which patients were most likely to respond — a biomarker approach that, if validated, would allow patient stratification rather than broad empirical trials.

For PLS specifically, there are two reasons ropinirole is worth considering. The first is the basal ganglia circuit argument made throughout this page: it targets the same motor loop that THC disrupts, through a different and enhancing mechanism. The second is the recognized subgroup of "PLS-plus" with parkinsonian features — bradykinesia, rigidity, postural instability — which represents a group where dopaminergic therapy has biological face validity. Published case reports (BMC Neurology, 2023) describe PLS patients who responded to levodopa, with variable but sometimes substantial benefit, particularly in the initiation and fluency of movement. This variable response makes sense if the degree of basal ganglia involvement differs between patients.

Practically: ropinirole is generic and widely available. The titration schedule from the ROPALS trial was cautious — starting at 0.25 mg three times daily and titrating upward over weeks toward a target of 16 mg per day — because dopamine agonists have meaningful side effect profiles at higher doses (orthostatic hypotension, nausea, somnolence, impulse control disorders). For PLS, where the primary goal is motor circuit enhancement rather than Parkinson's disease management, lower maintenance doses may be adequate, and starting slowly allows the patient and clinician to find the dose at which benefit is apparent without intolerable side effects. The ROPALS trial page covers the protocol, results, and limitations in more detail.

Levodopa/carbidopa — available in France as Modopar or Sinemet — is the alternative trial-of-therapy for basal ganglia engagement. It is the older, cheaper, and more immediately available option, and for a patient or neurologist uncertain about committing to a dopamine agonist, it provides faster feedback: levodopa's effects are often apparent within days to a couple of weeks at sufficient dose, which makes the trial-of-therapy more efficient as a diagnostic test of whether dopaminergic enhancement helps this particular patient. The published case reports of PLS-plus responding to levodopa used standard Parkinson's dosing (100–300 mg levodopa equivalent daily). Response is variable, which is consistent with the hypothesis that basal ganglia compensation is more prominent in some PLS patients than others.

Compounds that target reticulospinal drive

The brainstem reticulospinal system is the least understood of the three compensatory pathways considered here, and the pharmacological toolkit for directly targeting it is the thinnest. The descending reticulospinal tract originates in the reticular formation of the pons and medulla and projects to spinal interneurons and motor neurons, using serotonin and norepinephrine as principal descending neurotransmitters. In PLS, where the corticospinal tract is selectively damaged, the reticulospinal system is partially spared — its cells of origin are not Betz cells and not primarily targeted by the disease process. It is therefore a legitimate compensatory pathway, and one that a healthy motor system uses for posture, gait rhythm, and the automatic aspects of walking.

Serotonergic and noradrenergic compounds are the natural candidates. SSRIs and SNRIs affect descending serotonergic and noradrenergic tone, but their motor effects are indirect and modest — they are not meaningfully reticulospinal activators in any direct pharmacological sense. Their relevance here is primarily through mood: depression is common in PLS and is frequently undertreated. A patient who is depressed has reduced motor activation drive from frontal-limbic-motor connections, reduced motivation to exercise, and poorer engagement with rehabilitation. Treating depression in a PLS patient is always worth doing on its own terms, and if it secondarily improves walking motivation and effort, so much the better.

Droxidopa (L-DOPS, marketed as Northera) is a synthetic norepinephrine precursor approved for neurogenic orthostatic hypotension in conditions such as multiple system atrophy, pure autonomic failure, and Parkinson's disease with autonomic failure. Theoretically, by increasing systemic and central norepinephrine availability, it could support reticulospinal descending output. There is no clinical evidence in PLS and no MND trials. It is mentioned here for completeness, as a compound whose mechanism is directionally relevant even if the evidence is absent.

Atomoxetine, a selective norepinephrine reuptake inhibitor approved for ADHD, has been explored in small studies for its motor effects in other conditions, with inconsistent results. It is not a standard neurological intervention for motor impairment and has no PLS evidence. Like droxidopa, it belongs in the "mechanistically plausible, not clinically established" category.

This section is genuinely "watch this space." The reticulospinal system is an important compensatory pathway in PLS, but the pharmacological tools to directly and reliably enhance it do not yet exist in clinically validated form. The compounds above are not first-line targets. Cerebellar and basal ganglia circuit enhancement, described in the preceding sections, have meaningfully better clinical evidence and more clearly defined mechanistic rationale.

Approaches that would be mechanistically perfect but are not available

Intellectual honesty requires naming the pharmacological approach that would be most directly predicted by the THC observation — and why it cannot be used.

If the problem is acute CB1 activation causing cerebellar and basal ganglia motor dysfunction, the direct pharmacological inverse would be a CB1 antagonist or inverse agonist. Block or reverse the receptor activation, prevent the circuit suppression, maintain the compensation. In principle, a CB1 inverse agonist taken before activities requiring walking might protect against precisely the kind of circuit failure the THC observation revealed. It would not just be symptomatic — it would be directly targeted to the vulnerability the natural experiment identified.

There was a CB1 inverse agonist in clinical use: rimonabant (Acomplia), approved in Europe in 2006 for obesity. It worked for its metabolic indication. It was withdrawn from the market in 2008, approximately two years after approval, because of psychiatric side effects — specifically depression, anxiety, and suicidality — that emerged in post-marketing surveillance. These effects are not surprising in retrospect: CB1 receptors are expressed in limbic and prefrontal cortex circuits involved in mood regulation, and inverse agonism (which constitutively suppresses CB1 signaling rather than merely blocking agonist access) is a pharmacologically aggressive intervention in those circuits. Rimonabant is not available. It cannot be prescribed. No equivalent CB1 inverse agonist has been approved since.

Peripherally-restricted CB1 antagonists are being developed for metabolic and gastrointestinal indications, specifically engineered to have minimal central nervous system penetration. These avoid the psychiatric side effects of rimonabant but cannot, by design, reach the brain circuits where the relevant motor effects occur. They are the pharmacological equivalent of a solution to the wrong problem.

Two other compound classes are mechanistically interesting but not clinically available. FAAH inhibitors (fatty acid amide hydrolase inhibitors) modulate endogenous cannabinoid tone by preventing the enzymatic breakdown of anandamide, one of the brain's own CB1 ligands. Rather than directly activating or blocking CB1, they allow anandamide to accumulate at synapses where it is naturally released — a more physiological approach to modulating the system. Several FAAH inhibitors have entered clinical trials for pain, anxiety, and PTSD. However, BIA 10-2474, a FAAH inhibitor tested in a Phase 1 trial in Rennes in 2016, caused a catastrophic safety incident in which one participant died and several others suffered serious neurological injury. The mechanism was not fully clarified, but it triggered a complete reassessment of FAAH inhibitor development. No FAAH inhibitor is approved or ready for clinical use outside highly controlled trial settings.

MAGL inhibitors (monoacylglycerol lipase inhibitors) modulate the other major endocannabinoid, 2-AG, through a similar approach — preventing its breakdown and thus increasing its availability at CB1 synapses. Several are in clinical development for pain and inflammatory indications. None has MND evidence. None is approved. The story is similar to FAAH inhibitors: biologically interesting, pharmacologically plausible, clinically not available.

There is something genuinely frustrating about this section. The most mechanistically elegant approach to the circuit problem — a CB1 antagonist that protects the very compensatory circuits the THC observation revealed — does not exist in prescribable form. The class that existed was withdrawn for psychiatric reasons that are, in retrospect, predictable but were not prioritized. The next generation of targeted endocannabinoid pharmacology is at least five years from clinical availability, probably longer. In the meantime, the indirectly-targeted compounds in the preceding sections represent the practical frontier.

What to discuss with a neurologist

One practical observation about which neurologist to approach: a neurologist familiar with both motor neuron disease and multiple sclerosis — or one with a general movement disorders background — is more likely to engage productively with this reasoning than a clinician whose practice is exclusively ALS-focused. The Weill Cornell trial is an MND context, but dalfampridine's mechanism is cerebellar, and the framing of compensatory circuit targeting draws from the cerebellar and movement disorders literature as much as from MND research. A neurologist who has prescribed Fampyra for MS walking impairment and who understands the Kalla/Glasauer Purkinje cell mechanism will grasp the argument quickly. One whose only frame is the glutamate toxicity/neuroprotection paradigm of ALS research may find it less intuitive.

In France, the Salpêtrière MND center (Centre Référent SLA, Hôpital Pitié-Salpêtrière, Paris) and the network of regional Centres de Référence SLA are good starting points. Some of these centers have neurologists with MS and movement disorders backgrounds in addition to MND expertise, and the academic culture at Salpêtrière in particular is receptive to mechanistic reasoning.

Honest caveats

The reasoning on this page is mechanistically defensible. It is not established clinical practice. These are not the same thing, and the difference matters.

The CB1 receptor anatomy is real. The compensatory circuit hypothesis in PLS is real, in the sense that the disease clearly recruits non-corticospinal motor pathways as the primary pathway degrades — this is supported by imaging studies showing increased cerebellar and basal ganglia activation in PLS patients compared to controls during motor tasks. The inference that THC's acute motor disruption in this patient reflects heavy reliance on CB1-rich compensatory circuits is coherent. But no randomized trial has validated this specific framework as a basis for drug selection in PLS. The chain of reasoning has not been prospectively tested. It may be directionally right and still fail to produce clinical benefit in practice, because motor compensation in PLS is heterogeneous and the magnitude of any individual drug's effect may be below the threshold of detectability.

Individual response variability is substantial. The Weill Cornell dalfampridine trial in PLS enrolled 35 patients — already a small sample — and the responder pattern from the MS literature (40% responders) suggests that at least 60% of patients will not show clear walking improvement on dalfampridine even in the right disease context. The ROPALS trial used iPSC-derived motor neuron biomarkers to try to identify who would respond, which is clinically interesting but not yet a validated stratification tool that any neurologist outside Keio University can currently deploy.

Trial-of-therapy is therefore the right clinical approach for all of these compounds. The question is not whether to believe the mechanism but whether it produces a measurable functional benefit in this specific patient. Start at a reasonable dose, document the baseline objectively (Timed 25-Foot Walk, a patient-reported walking distance estimate, whatever is practical in the clinical setting), give the compound a defined trial period — two to four weeks for dalfampridine, four to eight weeks for ropinirole to reach an effective dose — and then make a decision based on the response, not on the theoretical rationale. The mechanism suggests the drug is worth trying. The outcome determines whether to continue.

All of these drugs are off-label in PLS. None is approved for primary lateral sclerosis. A neurologist who prescribes them is acting on mechanistic reasoning and clinical judgment, not on a registered indication. That is a legitimate thing for a neurologist to do, and it is what responsible off-label prescribing looks like — but it requires a neurologist who is comfortable with that kind of reasoning, and it is the patient's and family's job to bring the evidence to the conversation rather than expecting the neurologist to have assembled it independently.

Rapid deterioration in PLS deserves a workup before adding compounds. If the walking decline has accelerated, the first question is whether something treatable is driving it: B12 deficiency, copper deficiency myelopathy, thyroid dysfunction, new cervical myelopathy from spondylosis, or medication side effects (baclofen, tizanidine, and anticonvulsants can all worsen weakness at excessive doses). Repeat EMG should be considered to confirm the diagnosis remains pure UMN disease rather than conversion toward ALS with lower motor neuron involvement. The compounds on this page are for the patient where PLS diagnosis is confirmed and standard treatments are insufficient — they are not the first response to any acute change in function.

Finally: none of these compounds modifies the underlying disease. Dalfampridine, ropinirole, and acetazolamide target the compensatory circuits that are carrying the functional load, not the corticospinal tract damage that constitutes the disease itself. The effect of any of them, if present, is symptomatic — it enhances what is working, it does not repair what is damaged. This matters for how to interpret any improvement. A patient who walks better on dalfampridine has not experienced disease modification. They have had their remaining cerebellar compensation enhanced. When dalfampridine is stopped, the effect disappears. That is still worth having — "better walking while on this drug" is a meaningful outcome for someone with a progressive neurological disease — but it is important to be honest about what is being accomplished.

This page is not medical advice. The reasoning is shared to enable a better conversation between a patient, a family, and a neurologist who is willing to think about the specific physiology of the case.