Summary of a spinal dopaminergic and opioidic hypothesis of Restless Legs Syndrome

This is written for doctors, nurses, neuroscientists and other clinical and research specialists. 
Please don't make treatment decisions concerning drugs or other
medical interventions based
on this material or that on the main page
This material has not
been peer reviewed, though I have had encouraging feedback
from two neurologists. 

  Robin Whittle    2012-08-22, updated most recently 2020-05-30. 

2020-05-30 update:

I haven't worked on RLSD for some years, nor tried to keep up with the literature,  This was primarily because I found little new research concerning the spinal cord - which is obviously where the pathology is located.  It is as if many neuroscientists are uninterested in the lowly spinal cord when they could be working on the brain.

However, please see this 2019 article, which reviews and discusses all RLS/PLM work on spinal since the excellent 2000 article linked to below: ../#bara_jimenez Periodic limb movements in sleep - State-dependent excitability of the spinal flexor reflex

The 2019 review article is:

Restless legs syndrome: Clinical changes in nervous system excitability at the spinal cord level
Chloe Dafkin, Warrick McKinon and Samantha Kerr
Sleep Medicine Reviews Volume 47, October 2019, Pages 9-17 (Paywalled)

Please see the April 2019 article Spinal Cord Stimulation for Restless Legs Syndrome: Case Series and Mechanistic Hypothesis (Sci-Hub full PDF) regarding spinal cord stimulation leading to lasting relief for three patients who suffered from severe RLS, largely or entirely to spinal injuries.  This article is one of only a few to recognise the spinal etiology of RLS.  I suggest that the electrostimulation may increase blood flow in the relevant part of the spinal cord, which may also increase the delivery of tyrosine to the areas where (I suggest) it is depleted.  These surgical interventions are a welcome development, since as far as I know no amount of prescription drugs can provide good relief, and these always have severe ill-effects.  While this main page of this RLSD section contains several suggestions regarding salt, potassium, magnesium, caffeine, coffee, exercise and potentially percussive massage of the spinal cord, I doubt that these would be sufficient to provide adequate relief for the many patients with severe RLS due to generally unsolvable spinal injury problems.

The following is written as if opioidic inhibition of nociceptive neurons (those in the spinal cord which sense pain and drive the limb-withdrawal reflex action) occurs at their dendrites (inputs) as it does with dopaminergic inhibition.  However, the opioid receptors which inhibit the activation of these nociceptive neurons are actually on the neurons which drive them and inhibit the firing (depolarisation, with neurotransmitter release to the input dendrites of the nociceptive neuron) of these neurons.  The principle of inadequate opioidic and/or dopaminergic inhibition of nociceptive neurons remains the same.


I am an independent RLS/PLMD researcher in Daylesford Australia. 
I hope you will be interested to read this summary of my novel
observations, etiological hypotheses and my suggestions for
non-drug treatments:

Many researchers believe that the cause of RLS/PLMD is in the
brain.  However, there is no widely accepted etiological theory
for this common condition.  I suggest that the causes involve
inadequate dopaminergic and opioidic inhibition of a uniquely
human soft-touch-activated foot lifting and leg withdrawal
reflex circuit in the spinal cord. 

This reflex response is triggered by soft, non-painful, point-
contact touch in the foot arch.  The neurons which drive this
foot lifting and leg withdrawal are normally strongly inhibited
by their dopamine receptors being activated by a normally
continual supply of dopamine from the output terminals of a
descending dopaminergic pathway.  As with other neurons which
drive nociceptive reflex actions, there responsiveness is
reduced to the extent that their opioid receptors are activated.

Conventional foot lifting and limb withdrawal reflex responses
are driven by noxious stimuli: painful intrusion into the skin,
heat etc.  I propose that our foot-arch protective reflex
circuits have three special characteristics not found elsewhere
in the human body, or in any other mammalian species:

  1 - The response can be fully triggered by a non-painful
      stimulus which does not involve any damage to the skin:
      point-contact soft touch such as 1 gram as force, with
      the rounded tip of a pencil, the blunt end of a pencil,
      the head of a match or similar.  This sensitivity is
      only, or primarily, in the relative thin skin of the arch
      of the foot - not the heel, ball, toes, sides or top. 

  2 - In addition to being inhibited by their opioid receptors,
      the spinal neurons can also be inhibited by their
      dopamine receptors - which are unique to the particular
      nociceptive reflex neurons for this foot arch protective

  3 - These neurons' dopamine receptors are normally activated
      strongly since they synapse with the output terminals of
      neurons of a novel (evolved purely to regulate this
      reflex response) descending pathway, presumably from the
      brainstem or mid-brain or similar, which is normally
      activated and so continually releasing dopamine into
      synaptic cleft.  "Normally" means during sleep and most
      waking activities - all of which do not involve any risk
      of the foot arch being punctured by a sharp rock or
      spiny object.

      There must be some non-conscious control of this
      descending pathway which turns it off - and so enables
      the reflex response to be activated by the soft touch
      stimulus - when we are walking on a surface where such
      punctures of the foot arch skin might occur.  The
      response causes the foot to lift up, or at least not
      descend, onto the spine - before the spine punctures
      the skin.

The incidence of RLS/PLMD symptoms may also be affected by the
"noise levels" (random activation in the absence of stimulus)
of the point-contact soft-touch mechanoreceptors in the foot

The most striking aspect of RLS/PLMD is that it is movement
disorder whose symptoms are brought on by lack of movement. 
My explanation for this is that in the lower spinal cord,
where these dopaminergic output terminals are located, their
typically constant output can be greatly reduced by localised
depletion of tyrosine in the CSF in the immediate vicinity
of these output terminals - I guess a distance of 0.5 to 3mm. 

When this occurs to a degree which causes some or perhaps all
the reflex neurons to fire, in response to the "noise" levels
of firing by the sensory neurons, then the symptoms (disturbed
sensation, urge to move and involuntary movements) occur.

Another striking, but little-known, feature of RLS/PLMD is that
percussive massage of the lower spinal area (but not smooth
massage) can reduce or eliminate symptoms within seconds or
minutes, with the benefit lasting for tens of minutes to a few
(Here I am assuming that our observations and those of
some friends are representative of all RLS/PLMD sufferers.)

Physical activity - especially with a raised heart rate bringing
lots of tyrosine rich blood into the spinal cord's capillaries -
and/or percussive massage of the lower spinal region agitates
the CSF, redistributing tyrosine rich CSF from a few millimetres
away into the areas where it is needed to synthesize dopamine
in the output terminals of these descending neurons.  This
causes an increase in dopamine output, increasing the degree of
inhibition of the reflex neurons, ideally to the point where
their firing rate in response to the sensory neurons noise falls
to a rate low enough not to cause any of the RLS/PLMD symptoms.

I suggest that this is not a "disease", but the result of
recent evolutionary adaptations to bipedalism which do not
work very well.  The condition is best known by the sensory
movement diagnoses Restless Legs Syndrome and Period Limb
Movement Disorder.  However, I believe the condition, at
lower those required to meet these diagnostic criteria,
responsible for an even wider incidence of insomnia and

There are numerous ways of reducing the incidence of symptoms
without the use of drugs.  Low iron and magnesium levels are
well known to drive the symptoms, and any such deficiencies
should be corrected.  All these nutritional and lifestyle
related changes should be attempted before resorting to dopamine
agonists and opioids, both of which have serious long-term side

Most of the discussion below relates to people with occasional
or moderate symptoms.  It is my impression that most people
with the very worst symptoms have these in large part due to
spinal injury.  I believe that the dietary and other
suggestions mentioned in this website will greatly reduce or
eliminate symptoms for most people who have no spinal
injuries, and that these should be tried anyway for such
sufferer with spinal injuries to reduce the need for dopamine
agonists and opioids.

Some novel observations

My wife Tina and I - and some other people - find that RLS
symptoms and periodic limb movements - PLMs - can be reduced or
eliminated within seconds or minutes, with the
benefits lasting
typically for two hours or so, by percussive
massage of the
lumbar spinal region
.  This reduces or abolishes
symptoms in the
legs, whilst not affecting symptoms in the arms. 

In our experience, percussive massage of the upper spinal
region, just below the neck, similarly abolishes or reduces
symptoms in the arms, without affecting the symptoms in the
legs.  Again, the effects begin within seconds or a minute or
so, and last for two or so hours.

General rubbing and percussive massage of the whole body seems
to help as well, with effects which last for several hours -
which suggests that the massage produces endogenous opioids
which, like pharmaceutical opioid drugs, inhibit all
nociceptive neurons in the spinal cord, including those of the
foot-arch protective reflex.

Some sufferers have discovered exercises which provide
substantial relief from, or prevention of, RLS/PLMD symptoms -
relief far more extended than the few minutes of relief which
results from walking.  These exercises, including inversion,
vigorous pushups and "upside-down bicycling" (resting on the
shoulders and neck, moving legs in the air) all involve
stretching or other mechanical disturbances of the lower
spinal area.

Ingesting sugar also seems to have benefits which last for two
hours or so.  This would not be surprising if it could be shown
that for the particular individual, sugar ingestion boosts
levels of endogenous opioids.  This effect could also be
mediated by raised glucose levels improving many cellular

From mid-2014 to August 2016 we were taking a potassium
(potassium gluconate) equivalent to a quarter or
less of the
US RDA of 4.7 grams.  I think this has reduced my
low level of RLS/PLMD
significantly, and made the symptoms less
likely to appear when I
consume more salt (sodium) than usual. 
Tina thinks it has made no significant difference.
(Potassium nutrition is a fascinating and contentious field which I
hope to write about more on this site.  We now take about half the
US RDA, as four small drinks of 8.6% potassium gluconate over the day.
Together with lowered salt intake, it is our impression that this has
had significant benefits regarding generally health and lower anxiety.
People with kidney damage or some other conditions should not take
potassium supplements.  Anyone who does should ramp the dose up slowly
over a few weeks.  Contrary to the widely perceived notion that all
potassium salts have a very strong taste, potassium gluconate has a mild
taste - an 8.6% solution tastes about as strong as 0.5% sodium chloride.

We have not tried potassium gluconate as a short-term remedy for RLS/PLMD
since the important aspect of potassium nutrition is the intracellular
level, not the blood plasma level.  Increased potassium intake boosts
intracellular potassium in all cells of the body over a period of days or
weeks.  The potassium concentration in the plasma should generally remain
See: Richard D. Moore's 2001 book The High Blood Pressure Solution
and Charles Weber's book Potassium Nutrition.)

Many people find that lowering salt intake reduces their RLS/PLMD
symptoms.  We observe this regularly - salty evening meals or snacks
clearly do increase the prevalence of RLS/PLMD symptoms. 

It is not surprising that raising the potassium-sodium ratio
above the typically extremely low levels in most people's diet will
improve numerous aspects of health, not least by reducing the severity
of neurological conditions, since this ratio affects every cell in
the body, and affects the membrane potential (voltage) which is the
basis of all the functions of neurons. 

Some people report reducing RLS/PLMD symptoms with magnesium
supplements.  We do this but we don't have any short term
correlations which would constitute observations regarding the
effectiveness of magnesium supplementation.

We find that a gram or so of oral tyrosine reduces or eliminates
symptoms in the arms and legs within 10 to 15 minutes, with
benefits lasting several hours.  Some other sufferers have
discovered this independently, but I have not heard of doctors
suggesting it as a treatment. 

When RLS/PLMD symptoms are present, we find that the lightest
touch (but not firm rubbing) with a finger or small object such
as a pencil-tip on the foot arch (or palm of the hand) will
elicit an immediate spinal-reflex-like foot (or hand) withdrawal
response which is similar or identical to the PLMs.  The same
actions on other parts of the foot or hand have no such effect.
(See the W. Bara-Jimenez 2000 paper for an electrical equivalent of this
soft-touch experiment.)

Furthermore, we find that local partial anesthetic of the foot
arch is a still more powerful way of reducing or eliminating
symptoms in the feet.  In the past, if oral tyrosine and
percussive massage of the spinal area and the whole body fail to
stop Tina's RLS/PLMD symptoms, we are usually able to stop
(they are primarily or wholly in her feet and legs) by
70% clove bud oil (active ingredient presumably
eugenol) and 30%
coconut oil to the soles of her feet.  Now
she has better iron
levels and presumably higher magnesium and potassium
levels, her
symptoms are not so strong as to
require this.

I am not sure this is without negative consequences, so we only
did it when there is no other way of getting rid of the
Pharmaceutical local anesthetics such as EMLA (
lidocaine and
prilocaine) work too.  Clove oil works really well
and is less

Tina never gets symptoms in her hands now that she has improved
her iron status.  We only
discovered the local partial
anesthetic approach since her hand
symptoms ceased.  I assume
the clove oil is acting only on the
sensory neurons of the foot
arch rather than having a systemic
effect.  We haven't tested
this by applying it somewhere else.

We and other some people find that coffee, including decaf,
elicits or worsens RLS/PLMD symptoms
.  This is easily
in light of the widely recognized protective effects
of opioid
agonists and the little-known research, beginning in
1983, which
shows that coffee of all kinds, decaf or not,
significant quantities of mu opioid receptor

See a section below regarding caffeine, coffee and chocolate.

As many other sufferers have observed, both unusually heavy

exercise or little movement all day tends to cause or worsen
symptoms.  In the context of the following hypotheses, these
observations could be explained by the lack of exercise leading
to poor replenishment of spinal cord CSF tyrosine, due to little
blood flow and little mechanical disturbance of the CSF to
reduce the localised tyrosine deficit referred to above - and
heavy exercise leading to lower tyrosine levels there, due to
serum amino acids in general being diverted to unusually high
levels of muscle repair.

SSRIs are widely recognised as contributing to RLS/PLMD
symptoms.  See the serotonin section below for more on this.

Mark J. Buchfuhrer and Jill Gunzel also report that orgasm
is protective against RLS/PLMD symptoms, especially for
women.  "There are many anecdotal reports of sexual
stimulation and, especially orgasm, helping RLS symptoms. 
This appears to be more prevalent among female RLS
patients, who will often find that when they are not able
to fall asleep due to RLS symptoms, having an orgasm may be
the only measure that will alleviate their symptoms and
permit them to fall asleep."  This is easily explicable in
terms of increased levels of endogenous opioids inhibiting
the spinal reflex neurons mentioned above.

Key points of my etiological hypotheses

For brevity I have stated these as if they were facts.

  1 - We have a uniquely human soft-touch activated foot
      withdrawal spinal reflex response which protects the soft
      skin of the foot arch from puncture by withdrawing the
      foot before a spiny objects penetrates or damages the
      skin.  (Only humans have arched feet.)  When this reflex
      circuit is disinhibited, the foot withdrawal can be
      triggered by the slightest touch to the foot arch.  This
      is a typical nociceptive response except that it is
      triggered by a soft touch sensation which would not
      ordinarily be painful and that it can be inhibited by a
      novel descending dopaminergic pathway.  The activating
      sensation is slight pressure on a small patch of skin,
      such as is elicited by a light touch (a few grams as
      force) with the tip or the round end of a pencil.  Firm,
      broad, pressure or rubbing does not activate the reflex
      response - and may reduce the tendency of the response to
      occur.  (RLS/PLMD sufferers sometimes refer to themselves
      as "night walkers", since walking for hours at night is
      the only way they can find relief.)

      While this soft-touch activated foot withdrawal reflex
      has presumably evolved to protect the foot arch, it may
      also have evolved in part to protect the palm of the hand
      when climbing or grasping, or perhaps when walking on all
      fours.  Even assuming that it evolved purely to protect
      the foot arch, the genetic instructions which give rise
      to the requisite neuronal structures would probably also
      produce similar structures in other parts of the spinal
      cord, leading to similar soft-touch, opioidic and
      dopaminergic inhibited, muscle-activating (for limb
      withdrawal) reflex circuits for the palm of the hand, or
      for sensory neurons and muscles in the torso.  In extreme
      cases RLS/PLMD symptoms occur in feet, hands and torso -
      and there are a few reports in the literature of facial
      symptoms too.

      When I use a motorised massager on Tina's back, when she
      has RLS/PLMD symptoms, I usually start in the lower
      spinal area, then work on the upper part near the neck.
      After massaging her arms, I then go down the spine and
      on either side - vibrating her ribs robustly.  This
      frequently elicits spasms - but this only occurs when she
      has symptoms (in her feet) and the spasms reduce and
      stop after a few minutes as the effects of the massage
      reduce all of the symptoms in general.  (The reduction in
      the torso spasms could be caused by local desensitisation,
      increased blood circulation and tyrosine diffusion and/or
      general, whole-body or at least whole-spinal-cord
      endogenous opioid effects.)

      I propose that the percussive massage stimulates deeper
      mechanoreceptors driving spinal neurons which drive
      torso muscles - the same kind of circuit as with the foot
      and hand withdrawal reflex circuits.  There's no obvious
      benefit to having such muscle spasms, since the torso
      can't be withdrawn from a noxious stimulus. 

      I propose that these torso circuits, like those of the
      hand and arms, serve no evolutionary purpose, but result
      from the pattern of neuronal development which evolved to
      protect the foot arch being also expressed, to a lesser
      degree, in other segments of the spinal cord - there being
      no particular evolutionary impetus to develop a more
      targeted developmental system.

  2 - Most of the time - including while sleeping and when not
      walking on potentially spiny surfaces - the spinal neurons
      which mediate this foot
arch protective response (or
      similar neurons for the foot and torso) are inhibited by a
      uniquely human set
of dopaminergic (dopamine-producing)
      neurons.  These
presumably descend from the brain or
      brainstem and are not under conscious control. 

      These spinal neurons, like all the other nociceptive and
      withdrawal reflex generating neurons in the spinal cord
      (the same neurons do both functions), are also inhibited
      to some extent by opioid receptor activation, including
      that which results from the normal moderate levels of
      endogenous opioids in the CSF of the spinal cord or
      elevated levels due to exercise, sex or perhaps the
      experience or anticipation of particularly painful

  3 - The descending neurons release dopamine most of the time -
      whenever we are not walking on ground which might contain
      foot-arch puncturing spines, spikes or rocks.  These
      neurons require a nearly continual supply of tyrosine, in
      order to produce
L-DOPA which is then converted into
      dopamine.  While these neurons presumably have their
      nuclei in the brain or brain stem, their output terminals
      are in the lower spinal cord, and the tyrosine must be
      produced in those output terminals - so this is where the
      tyrosine must be taken from the CSF and brought across the
      neuronal membrane into the output terminal.

  4 - The tyrosine supply to these neurons (from the spinal
      cord's CSF) can become depleted at times of lowered
      physical activity, since without vibration and/or high
      rates of blood-flow, the consumption exceeds that
to the CSF by the immediately adjacent
      capillaries and the
natural degree of diffusion of
tyrosine through the CSF in this dense part of the spinal
      cord (the lumbar region, for the circuits controlling the
      foot and leg muscles). 
Following an hour or more of
      physical inactivity, this results in
a localized depletion
      of tyrosine, such as within one or two
millimetres of the
      output terminals of these spinal
dopaminergic neurons,
      which are presumably in the dorsal
horn.  When this causes
      insufficient dopamine release to
properly inhibit the
soft-touch-activated foot-arch protective reflex circuit,
and/or PLMD symptoms occur - due to low levels of
      firing by the sensory neurons, even in the absence of any

      This would explain the puzzling characteristic of RLS/
being a movement disorder which is triggered or
by lack of movement.

      Conventional research, which involves non-human animals
      indicates that there is little or no dopaminergic
      activity in the mammalian spinal cord.  The recently
      evolved (~3 million years since the evolution of
      bipedalism and so our foot arches) nature of this
      soft-touch activated reflex response and its normally-on
      dopaminergic descending pathway means that humans have not
      had much time to develop a better tyrosine supply for
      these neurons' output terminals.  Furthermore the great
      demands on the lower spinal cord due to bipedalism means
      that the lumbar cord as grown in complexity, size and
      perhaps density in the last few million years.  So humans
      probably have unique difficulties with blood and CSV
      circulation in the lumbar area. 

  5 - Robust physical movement, spinal stretch or inversion
      exercises and especially percussive massage shakes up the
      CSF in this area, bringing tyrosine-rich CSF to these
      neurons from a few millimetres away, within seconds or
      minutes.  This restores higher
levels of dopamine
production and release, which more effectively inhibits
      the foot-arch protective soft-touch activated spinal
reflex circuit, leading to a reduction or cessation of

  6 - With very low levels of inhibition (due to lower levels
      of activation of the opioid and dopamine receptors of the
      nociceptive/reflex neurons) this circuit can be triggered

      by (what I assume is) the low "background noise" level of
      foot-arch soft-touch sensory neuron activation which
      occurs in
the absence of any stimulation, leading to RLS
      sensations and to the PLMD foot (or hand) withdrawal
      actions - or in the torso, to muscle spasms even in the
      absence of any touch or deeper stimulus, such as that of
      the motorised percussive massager.  I know of no
      that RLS/PLMD symptoms
arise from any fundamental problems
      with the
foot-arch skin soft-touch sensory neurons
      themselves.  However, see the serotonin section below. 

      "Background noise" is an audio or radio engineering term.
      I assume that sensory neurons, even if not stimulated at
      all, will occasionally fire of their own accord.  I have
      not read of this in the literature and would especially
      appreciate someone pointing me to any relevant references.

  7 - From the above (and assuming that the "noise level" of
      soft-touch sensory neuron activation remains the same
      (which is not necessarily the case, as mentioned below in
      the serotonin section) it follows that the pathological
      state in
which RLS/PLMD symptoms are present is due to
      this extreme
state of disinhibition of the uniquely human
      spinal reflex
circuit allowing the noise level of sensory
      neuron activity causing firing of the reflex neurons to a
      degree necessary to result in one or both sets of
      perceptible or observable symptoms:

         a - The sensory feeling of restlessness, which may
             be immediately and very temporarily relieved by
             moving or shaking the foot, or rubbing it, or by
             walking.  (This is the RLS subset of the
             condition's total symptomatology.)

         b - Involuntary movements: the distinctive toe
             uplifting movements, typically a few such
             pulses in a second or two, with such bursts
             repeating every 10 to 30 seconds or so; full limb
             withdrawal movements, perhaps triggered by the
             toe lifting reflex action or directly from the
             reflex neurons themselves.  (This is the PLMD
             subset of the condition's total symptomatology.)
ssuming there are no congenital, injury, disease or
      developmental problems with the spinal circuitry, a
s far
      as I know, the causes of
this high degree of disinhibition
      fall into four categories.  For simplicity in the
      following descriptions I also assume that the numbers and
      efficiencies of the opioid and dopamine receptors in the
      spinal reflex neurons are unaltered - but such variations
      may indeed exist and so contribute to the level of

      Categories 1
and 2 apply to all spinal nociceptive/reflex
      circuits.  Categories  3 and 4
only apply to those
circuits which are, or which resemble (for the hands and
torso, where similar symptoms may also occur), the
recently evolved (last ~3 million years) human soft-touch
activated foot-arch protective reflex circuit.

        1 - Any compounds which inhibit the mu opioid receptors
            in this part of the spinal cord.  This includes some
drugs and the compounds in coffee such as 
            4-Caffeoyl-1, 5-quinide.

        2 - Lower than normal levels of endogenous opioids - or
            levels which are unable to compensate for the
            effects of the mu opioid receptor antagonists.
        3 - Similarly, any compounds which antagonise the
            dopamine receptors in these spinal neurons, such as
            promethazine (Phenergan).   

        4 - Similarly, low levels of dopamine receptor agonists
            in the receptors of these foot-arch reflex spinal
            neurons.  Without drug intervention, the only such
            agonists are the dopamine molecules released by the
            descending dopaminergic neurons.  The most
causes of these low levels of dopamine
            release are:

               a - Iron deficiencies in general.  (Iron is
                   essential for the functioning of tyrosine
                   hydroxylase, which converts tyrosine to
                   L-DOPA.)  There may also be problems with
                   the ferritin/serum iron balance and CSF
                   iron levels.  Low iron levels have long been
                   recognised as a cause of RLS/PLMD.

               b - Iron deficiencies in these spinal
                   dopaminergic neurons.  Perhaps some people
                   have a genetic predisposition to reduced
transport across the blood brain
                   barrier and/or from the CSF to the cytoplasm
                   of the
neurons.  This would would have been
                   adaptive in
ancestral times where high-iron
                   diets or water
supplies might otherwise lead
                   to neurotoxicity.  (I recall some research
                   which involved a genetic predisposition to
                   RLS/PLMD symptoms due to a mutation in a
                   gene for a protein which was involved in
                   iron transport into cells.)

               c - Anything else which reduces the activity of
                   tyrosine hydroxylase in these neurons, such
                   as genetic variations in the gene for this
                   enzyme, enzymes involved in its co-factors
(perhaps) reduced folate levels.

               d - Low levels of tyrosine in the blood and so
the CSF.  This could be caused by
rhythms, low protein levels in
                   recent food
intake or high levels of
                   exercise causing
greater than usual amino
acid consumption as muscles are rebuilt.

               e - Potential problems with transporting
across the blood-brain barrier (in
                   the spinal
cord) and similarly into the
neurons as noted in b above. 
For instance higher plasma levels of the
                   large neutral
amino acids other than of
tyrosine, since these compete with tyrosine
in the Large Neutral Amino Acid transporter
system.  (I am keen to find a good description of
                      this this system.)

               f - Localised (within a millimetre or two)
                   depletion of tyrosine due to recent lack of
                   physical activity and vibration as explained

               g - Additional localized shortage of tyrosine
other nutrients in the general area of
dopaminergic neurons (millimetres and
                   centimetres) due to the effects of spinal
                   injury, including that resulting from spinal
                   surgery.  (Some people who have never had
                   RLS/PLMD develop it on the first night after
                   spinal surgery or injury - and the symptoms
                   remain indefinitely.)

                   I assume that the spinal injury disrupts the
                   flow of blood and/or CSF - so that in the
                   small volumes (I guess a few tens or at most
                   hundreds of cubic millimetres) where tyrosine
                   is continually consumed, the injury reduces
                   the ability of the blood to replenish CSF
                   tyrosine and/or the ability of nearby
                   tyrosine-rich CSF to diffuse that tyrosine
                   into the depleted area.
  8 - In this model, it is easy to see why both dopamine
      agonists and opioid agonists are highly effective at
      reducing or eliminating RLS/PLMD symptoms.  However, both
      types of drugs have serious side effects, and I suggest a
      variety of non-drug treatments be tried before resorting 
      to these interventions:

        a - Improve iron intake under medical supervision (as
well known).

        b - Eliminating all coffee, or at least avoid it in the
            afternoon and evening.

        c - Eliminating caffeine.  See my notes below on how it
            may reduce RLS/PLMD symptoms in the 2 to 4 hours
            after ingestion, but worsen the symptoms 5 or more
            hours after ingestion.  Caffeine is also a sleep
            disruptor and drives anxiety, irritability and
            tiredness (except the first few hours).

        d - Reduced salt and improved magnesium and calcium.  I
            don't know the mechanisms but we and some other
            people find that excessive salt causes or
            exacerbates RLS/PLMD.  Some people find that
            improved intake of magnesium or magnesium and
            calcium reduces or eliminates their symptoms.

            (Sodium is not necessarily the only cause of
            ill-health due to excessive salt consumption -
            the chloride could be involved as well.  We
            generally only get substantial sodium or chloride
            via salt, so we tend to get the two together.)
        e - Improving potassium nutrition.  See notes below
            about potassium supplements.  In the absence of
            supplements, eat more bananas, potatoes and other
            strong sources of potassium which are low in sodium.

        f - Limiting or avoiding chocolate and cocoa, including
            raw (unroasted) cacao.  We don't know the mechanism,
            but we find, unfortunately, that these drive
            RLS/PLMD symptoms.

        g - Eating high protein meals before bed.  (There may
            be reasons to avoid this, not least the desire to
            boost serotonin in general, or at least at night:
   .  Low serotonin
            levels are widely recognised as a cause of mood and
            other problems.  Unfortunately, as noted below in
            the serotonin section boosting serotonin levels may
            worsen RLS to some degree.)

        h - Ensuring that the arms, shoulder and especially
            the back are kept warm when sleeping.  See point
            9 below.

        i - Exercise and/or percussive massage of the spinal
            area (if safe) before or during the times when
            symptoms are present.

        j - Oral tyrosine, such as 2 or 3 500mg tablets or
            capsules.  Tyrosine appears to be safe (see the main
            page ../ for references).  It is  rapidly ingested
but it is poorly soluble.  Ideally there would be
slow-release, high availability, tyrosine tablets. 
            Free amino acids such as
tyrosine are arguably
            nutrients, rather than drugs.

        k - General massage of the whole body, wherever safe,
            including percussive massage and rubbing, and
            including the soles of the feet.  I am not sure why
            this is so effective, but I suspect it is probably
            due to release of endogenous opioids and/or other
            DNIC (Diffuse Noxious Inhibitory Controls)
            mechanisms which reduce the sensitivity of all
            spinal nociceptive neurons.

            I suspect (I don't know of any research evidence)
            that stretching exercises also boost the levels of
            endogenous opioids, so this may help reduce RLS/PLMD
            symptoms too.

        l - Local partial anesthesia of the foot arch and sole,
            and perhaps the palms of the hands.  This is a drug
            approach with potential negative side effects, but
            perhaps it would be preferable to the systemic
            administration of opioid agonists and dopamine

        m - Avoid antihistamines or other drugs which, in the
            sufferer's experience, cause or worsen RLS/PLMD

  9 - There is an interesting and consistent observation we
      have both made of ourselves, but have not yet compared
      notes with anyone else regarding this: when sleeping or
      trying to  sleep, in a condition where RLS/PLMD symptoms
      are on the edge of occurring, we find that if our arms,
      shoulders or back are cold, such as being uncovered or
      exposed to draft, then the symptoms will occur. 

      Likewise, we find that lying on our back will typically
      cause the symptoms to occur, while lying on our side does
      not.  (I think one correspondent reported the opposite -
      symptoms worsened when lying on their side, and reduced
      when lying on their back.)  At first I thought the back
      pressure and temperature effects might be due to blood or
      perhaps CNS circulation changes, but now I think it is a
      direct sensory stimulus process on the skin, since the
      effect on symptoms seems to occur within seconds.

      I discussed these observations in the PDF file I wrote for
      the main RLS researchers in 2011.  These are of practical
      use in reducing symptoms, and potentially of interest to
      researchers concerned with CNS development.  I propose
      that cold, or pressure, in these areas is driving the
      same neurons in the lower spinal cord which are primarily
      driven by the soft-touch-based (as well as pinprick-pain-
      based) foot-arch sensory neurons.  I propose that these
      observations reflect the existence of inputs to this
      circuit from other parts of the body which were not
      pruned during early (in utero, I guess) development. 

      Since cold sensations are processed by nociceptive
      circuits in the spinal cord, similarly to the signals
      from pain receptors which report tissue damage, and since
      the particular circuit for the foot arch has evolved to
      respond to soft touch as well, such crosstalk from cold
      and soft pressure
sensory neurons in other part of the
would not be surprising - to the degree that the
      axons from these sensory neurons to the foot-arch neurons
      were not completely pruned.

      My primary reference for understanding this crosstalk is:

         Appropriate/Inappropriate Developed Pain Paths
Jens Schouenborg     
         Chapter 12, in The Science of Pain (1056 pages)
         Alan Basbaum and M. Bushnell eds
         Elsevier Academic Press Aug 2008 
         ISBN 978-0-12-374625-2
      I found chapters 23 (Morphological and Neurochemical
      Organization of the Spinal Dorsal Horn), 24(Spinal Cord
      Physiology of Nociception), 25 (What is a Wide-Dynamic-
      Range Cell), 49 (Descending Control Mechanisms) and 50
      (Diffuse Noxious Inhibitory Controls (DNIC)) really
      useful too.

 10 - My proposal of a mechanism involving a localized tyrosine
      deficiency in the spinal cord is unlike any neurological
      condition I have read of.  However, it may have a
      parallel with foot cramps.  I find that these can often
      be resolved, in a few seconds or tens of seconds, much
      more rapidly than by rubbing the foot or calf, by
      percussive massage of the lower spine, exactly as for
      RLS/PLMD.  This doesn't require a mechanical massager -
      I simply stand up and use my fists or the backs of my
      hands on either side of my lower back.

      So perhaps some foot cramps involve either a localized
      deficiency of some precursor(s) and/or a localized build-
      up of reaction byproduct(s) which inhibit a particular
      lower motor neuron process.  Shaking up the cord and its
      CSF seems to fix any such problems.

      There is a well known co-morbidity between foot/leg
      cramps and RLS/PLMD.  Perhaps the common factor is poor
      circulation of the blood and CSF in the spinal cord -
      even if the precursor which is presumably deficient for
      foot cramps is not tyrosine.  This would be the case for
      those people with spinal injuries - and not everyone who
      has a spinal injury is aware of it.
      I have not yet tried to find whatever research literature
      they may be on foot cramps.

My hypotheses suggest that multiple systems can contribute to
the disinhibition of the foot-arch protective reflex response,
so it is not surprising that a large number of factors seem to
affect the level of symptoms. 

The central principle of my hypotheses is that RLS/PLMD
only arise due to an arguably pathological degree of
disinhibition of what I propose is a uniquely human spinal
reflex circuit, which evolved to protect the soft skin of our
foot arch, and perhaps of the palm of our hands.  This pattern
of neuronal development is evidently present not just for the
foot arch, but for the palm of the hands and for some torso
sensory and motor neurons.  In extreme cases (low iron, high
sodium, high levels of opioid receptor antagonists such as
those in coffee, maybe dopamine receptor antagonists, low
protein levels, poor circulation etc.), all three systems

can be triggered by "background noise" of the sensory neurons,
leading to RLS sensory symptoms and PLMs in all three parts of
the body.  (There have also been reports of RLS/PLMD-like
symptoms in the face.  I haven't researched these but perhaps
the genetic instructions which give rise to the adaptive soft-
touch sensitive foot-arch system also create similar structures
at all levels of the spinal cord.)

As far as I know, there's nothing which could go wrong solely
the brain which would cause this disinhibition or the same
pattern of symptoms, which always (as far as I know) are reduced
by oral
tyrosine, increased blood-flow within and/or vibration
of the
spinal cord, increases in opioid and dopamine receptor
and reduction in any level of opioid and dopamine
antagonist activity.  For instance, if the (presumably
spinal) drive for the descending dopaminergic pathway was
functioning, this would disinhibit the soft-touch foot-arch
protective reflex circuit, giving rise to symptoms.  But in
such a condition, oral tyrosine would have little or no effect.
would percussive massage of the lower spinal cord area,
it could be shown that this in some way aided opioid
- which I think is possible.

While these processes are pathological, in that they cause
distressing symptoms, they are also very common.  I suggest
is due to the recently evolved nature of the proposed
foot-arch protective reflex, and the unusually high
demand for
tyrosine it imposes on certain narrow columns within
the spinal
cord.  The human body plan has not yet evolved a
sufficiently good blood supply to the lower spinal cord, since
the development of bipedalism has greatly increased the size
and complexity of the neuronal circuitry there - and the
selection pressures due to RLS/PLMD symptoms have not been
strong enough in the time available (~3 million years) for
better tyrosine supply arrangements to evolve and spread to all

Beyond the widespread symptoms which meet RLS/PLMD diagnostic
criteria, I propose that there is an even more widespread
occurrence of the same processes, to lesser degree, which
manifest as sleep disruption and difficulty getting to sleep. 
If so, then this process contributes enormously to the global
burden of insomnia, sleep disturbance and so to general
ill-health - not least of which are vehicle accidents due to
tiredness.  Also, any sleep disruption leads to people being
more likely to use caffeine, and then to the anxiety and other
ill effects which follow, including the use of alcohol and other
drugs to quell the anxiogenic effects of the caffeine.

It is my impression that those who suffer the most from RLS/PLMD
do so due to spinal injuries and/or due to coffee and other
aspects of nutrition and lifestyle which are generally amenable
to change.  Genetic disposition is no-doubt a factor but I think
the most severe cases often involve spinal injury - including
pressure or changes to blood / CSF circulation after otherwise
successful spinal surgery.

I think there is great scope for reducing or eliminating
symptoms without the use of dopaminergic and opioidic drugs, all
of which have serious negative consequences, not least with the
dopamine agonists, gambling addiction (link) and other
personality changes.  This will require research funding from
somewhere other than drug companies.

Serotonin - new research in 2016
SSRIs are well known to cause or exacerbate RLS/PLMD symptoms,
for some people at least.  Haba-Rubio et al. 2014
(Sleep Journal) report that SSRIs, but not tricyclic
antidepressants, worsen symptoms.

Since I was not aware how serotonin might be involved in
nociception, I wondered whether it might affect the random
"noise" rate of activation, in the absence of stimuli, for the
foot arch soft-touch neurons.

I think this article supports the idea that increased serotonin
levels in the foot-arch itself may contribute to this level of

    Merkel disc is a serotonergic synapse in the epidermis for
    transmitting tactile signals in mammals
    Weipang Chang, Hirosato Kanda, Ryo Ikeda, Jennifer Ling,
    Jennifer J. DeBerry, and Jianguo G. Gu
    PNAS vol. 113 no. 37 2016-10-11 

This is assuming that Merkel disks or something like them are
found in the foot arch, which is something I have not
established.  The lead author is quoted, in the press release:

   "The serotonergic transmission in the epidermis,
    probably like that in the central nervous system,
    can be regulated by factors affecting serotonin
    uptake and release,"

Assuming that SSRIs affect serotonin reuptake transporters in
soft-touch sensory neurons in the foot arch, and that the
"noise" of their activation in the absence of physical stimulus
is a key component of RLS/PLMS symptoms, then it might be
possible to apply a topical drug there to reduce serotonin
levels, and so make these neurons less likely to fire
sponteneously.  If this could be done in a way which did not
upset serotonin levels elsewhere in the body, then this might
be useful therapeutic technique.  (If I had the money and desire to
invest in patents, I wouldn't be writing this in public.)

This suggests a simple experiment: with a subject prone to
RLS/PLMD (easy to find, there's usually one or a few in every
ten people) apply solution with alcohol of fluoxetine or some
other SSRI topically to one foot arch with the other as a
control (apply a similarly textured liquid there) and see what
happens.  Does RLS/PLMD appear now in the SSRI treated foot?
Is the treated foot arch more sensitive to being touched
lightly with the tip of a pencil?

Conversely, does topical application of a serotonin antagonist
to the foot arch reduce or abolish RLS/PLMD symptoms?  It would
be best to avoid anything which was also a dopamine or opioid
antagonist as well.

Caffeine, coffee and chocolate
Caffeine is widely regarded as a causative factor for RLS/PLMD.
In 2015 I wrote, in a letter to the Journal of Caffeine
Research that I did not know of any research showing that
caffeine contributes to RLS/PLMD symptoms.  I was wrong to
state this, for at least the two following reasons. 

Firstly, in a book chapter I should have been aware of,
"Nonpharmacological Considerations and Treatment of Restless
Legs Syndrome", Mark J. Buchfuhrer (link) and Jill Gunzel cite
1978 research in which RLS/PLMD was eliminated by cessation of
caffeine.  The text of the article is here and the references
are here.  This is Chapter 30 (link) of their 2009 book
"Restless Legs Syndrome A volume in Pergamon Texts in Inorganic

The cited 1978 article is not in the archives of the
Journal of Clinical Psychiatry, but the abstract is at:

Restless legs, anxiety and caffeinism.
Lutz EG.
J Clin Psychiatry. 1978 Sep;39(9):693-8.

Secondly, it is our impression that for some sufferers at
least, caffeine can reduce or eliminate symptoms immediately
and for to 2 to 4 four hours or so after ingestion - and that
caffeine ingestion can increase RLS/PLMD symptoms in time
periods such a 5 hours and longer after ingestion. 

I am not claiming this is a solid observation or that that it
applies to all RLS/PLMD sufferers.  However, you don't need a
doctor's instruction or a pile of peer-reviewed journal articles
to experiment with caffeine and see how it affects your own
condition.  If this observation is valid and applicable to a
particular person, then we might find that:

  1 - Caffeine withdrawal increases RLS/PLMD symptoms.

  2 - Caffeine in the short term reduces those symptoms - so
      sufferers may regard the short-term effect as more
      important then the long-term effect (see point 1) and
      conclude that caffeine either doesn't cause symptoms or
      that it overall reduces them.

  3 - Perhaps caffeine in the short term - say a few hours -
      does reduce RLS/PLMD symptoms in someone who has not used
      any for a week or so.  This is hardly useful, since the
      symptoms can appear more frequently than this, and regular
      use of caffeine puts the user in a continual state of
      withdrawal, only alleviated briefly and partially for a
      few hours after each dose of caffeine.
Buchfuhrer is arguably the world's most foremost RLS
specialist doctor.  The late Jill Gunzel (obituary) was the
author of "Restless leg syndrome: the RLS rebel's survival
guide" in 2006 (Goodreads).  The fact that they caution against
caffeine suggests that it probably does contribute to RLS/PLMD
    "Although most RLS patients will note worsening of
     their symptoms with caffeine, there are no other
     reports in the literature demonstrating complete
     resolution of RLS complaints following the
     cessation of caffeine intake."

They also state that anxiety worsens symptoms, and caffeine
directly increases anxiety, irritability etc.  It is also a
sleep disruptor.

Likewise they warn against alcohol, which is a sleep disruptor.
Just because alcohol may have a short-term sleep inducing and
relaxing effect doesn't mean that there are any benefits in
using it continually.

Contrary to the above, some 2016 research shows no significant
relation between caffeine use and RLS/PLMD:

   Lifestyle Factors and Risk of Restless Legs Syndrome:
   Prospective Cohort Study.
   Batool-Anwar S, Li Y, De Vito K, Malhotra A, Winkelman J5
   Gao X.
   J Clin Sleep Med. 2016 Feb;12(2):187-94.

They cite the 1978 Lunz research and state of their own work:
"To our knowledge this is the first prospective study examining
the link between caffeine and the risk of RLS."

My view is that there are numerous reasons for reducing
caffeine consumption or avoiding it entirely.  It is a sleep
disrupting, anxiogenic, tiredness-inducing, highly addictive
drug.  It would not be surprising if it directly, or indirectly
via increased stress and disturbed sleep, causes or worsens
RLS/PLMD symptoms.

With great regret we observe repeatedly that chocolate also
drives RLS/PLMD symptoms.  Milk chocolate in small
quantities is not a problem, but dark chocolate, cocoa drinks
and the like definitely do increase the risk of symptoms.
The effects seem to last quite a while, in that dark
chocolate consumed in the early or mid afternoon can drive
symptoms 12 to 20 hours later.

The causative factor is unknown and is presumably not due
to opioid receptor agonist activity, since the researchers
who found this in coffee in 1983 (as mentioned above) did
not find such activity in chocolate. 

I suspect that the time frame for coffee's opioid antagonist
effects on RLS/PLMD symptoms is shorter than that of
chocolate, which would be further evidence that the
chocolate mechanism is different, or at least involves
currently unrecognised opioid receptor antagonists which
have a longer half-life in the body than those found in

Caffeinated coffee may have a short-term effect of reducing
RLS/PLMD symptoms, as some people report, due to the caffeine
either doing so directly, or for regular users, providing
temporary respite from the caffeine withdrawal condition which
makes RLS/PLMD symptoms worse.  In other words, the short term
beneficial effect of caffeine may be more significant than the
ill effects of the opioid antagonists.

The RLS researchers were informed about these observations and hypotheses in 2011
I wrote an earlier and much more detailed and well-referenced
version of this material (except for our recent discoveries
potassium the 2016 material on serotonin and the 2017
material on caffeine) in a 32 page PDF document and sent it to
the main RLS
researchers in September 2011.  I haven't received
any feedback
from them yet. 

These researchers are busy doctors and academics with a huge
amount to read every week to keep abreast of their fields. 
naturally do not wish to be associated with flaky theories
and there are plenty of such theories regarding RLS.  They
probably not very interested in reading at length what an
electronic technician
from Australia is doing in the field they
have been
researching for decades.  Yet all this professional
RLS/PLMD research has so far failed to provide even a candidate
comprehensive etiological theory - much less one which is
widely accepted.  (For instance, the 2016 article mentioned
above states: "The etiology of RLS is not known;".

Electronic technicians get lots of practice debugging systems
which are functioning differently to how they were designed to
work and/or which were never designed properly in the first
place.  Likewise computer programmers, working with complex
not very well documented circumstances in which something is
going wrong, and we need to find out all about it and fix it.

We frequently have to work without circuit diagrams or
adequately documented source code with little or no initial
knowledge of what should be occurring and what actually is
going on.  Many electronic faults involve inadequate power
supply, which is why it was easy for me to think of localised
depletion of tyrosine, the precursor to L-DOPA.  Likewise noise
(random fluctuations) are frequently a problem in electronic
equipment.  So it is easy for me to imagine low levels of
sensory neuron activity becoming a problem if the super-
sensitive neuronal circuit is disinhibited.  I have not read
of any currently recognised neurological conditions involving
such mechanisms - and in 2004/5 I did read the whole of Kandel's
Principles of Neural Science, 4th ed. 

The first discovery Tina and I made
- the effectiveness of
percussive massage of the
lumbar spinal area - occurred because
for an electronic technician it is standard procedure to
vibrate, heat, cool, bend etc. various parts of a piece of
equipment which is not functioning, or which is exhibiting
intermittently faulty behaviour.  This will frequently alter the
behaviour and so enable the discovery of a bad solder joint, an
overheated or degraded transistor etc.  (Heating the lumbar area
seems to reduce symptoms too, but not as well as percussive
massage - and I am concerned about overheating any part of the
body, especially the CNS.)  Our most recent tentative discovery
- of the importance of potassium - had nothing to do with me or
our interest in RLS: it was due to Tina's broad interest in
nutrition and health.

RLS/PLMD is extremely common, easy to research and currently
lacks an accepted etiological theory.  In the future I plan to
write a page suggesting the numerous experiments which can be
done, such as with percussive massage, oral tyrosine, the soft-
touch test of the foot arch, local anesthetic on the foot-arch,
exercise techniques etc. 
This research requires no drugs or
special equipment and may involve few safety concerns.  These
would make excellent undergrad,
post-grad and doctoral research
projects - and would be a valuable addition to the drug tests,
genetic research and assumptions about this being a brain
disorder, in which mainstream RLS/PLMD research has been stuck
for decades.

Since late 2011, has been a less
detailed, publicly accessible and updated
version of the
material I wrote in the PDF for researchers. 
Since then I have
been corresponding and talking with a number
of sufferers and
have have received appreciative responses from
two neurologists
- in Australia and the USA.  Several friends regularly use
percussive massage, with their own fists, if they get a bout of
RLS/PLMD.  One friend avoids coffee in the evening and uses
percussive massage and/or oral tyrosine if she has any symptoms.

I suggest that the condition be known not as a disease or
syndrome, but as
Restless Legs Sensorimotor Disorder.  For my
arguments against the name change from "Restless Legs Syndrome"
to "Willis-Ekbom Disease" and my concerns about this condition
being listed in the DSM-5 (it is a movement disorder, not a
psychological disorder), please see:

Update history
To the index page of this site: ../../
Contact details and copyright information: ../../contact/
© 2011-2020 Robin Whittle   Daylesford Australia