Analysis of parkinsonian tremor and rigidity relying on peripheral and central intensifying mechanisms through thalamic activities

Authors


Abstract

Background and Aim

Passive and active muscle manipulations modify parkinsonian tremor and rigidity. We hypothesized that these outer maneuvers provoke thalamic 3 to 7-Hz (τ-range) and 13 to 27-Hz (β-band) activities, which in turn act to develop and maintain the symptoms. To test the hypothesis, we examined the temporal and quantitative relationships of the thalamic activities with the symptoms in conditions under the outer maneuvers.

Methods

In thalamotomy for 20 patients with Parkinson's disease, we monitored thalamic local field potentials and multiple unit spikes with surface electromyograms of neck and limb muscles.

Results

Tremor is intensified in postural and kinetic modes with higher amplitude and faster frequency than the resting state, and in tremor rhythms on the cogwheel rigidity with accelerated muscle tone. Exaggerated thalamic τ-range activities always led to tremor of those intensified modes. The τ-range activities developed and maintained tremor in keeping a quantitative relationship in amplitude. Both the passive and active muscle manipulations transiently abolished the β-band activities, and then made them relapse to lead to rigidity in a waning and waxing fashion. Because of these changes of β-band activities, the patient's active effort to move or hold postures caused a state in conflict between voluntary and involuntary contractions.

Conclusion

Together, the results suggest that passively activated peripheral kinesthetic afferents, and actively triggered central motor initiatives act to develop and maintain tremor and rigidity through the thalamic motor nuclei as the crucial relay stations.

Introduction

In thalamotomy targeting the nucleus ventralis lateralis (VL) and nucleus ventralis intermedius (VIM), we observed a variety of expression of intensified parkinsonian tremor and rigidity in the routine clinical test of passive and active muscle manipulations. These outer maneuvers possibly drive the peripheral kinesthetic afferents and central motor initiatives. It is plausible that, though tremor and rigidity are centrally originated,[1, 2] their development and maintenance rely on some central and peripheral intensifying mechanisms. As the thalamic activities in and around the surgical targets are most susceptible to the outer muscle maneuvers, we analyzed the thalamic responses to test the hypothesis that the intensifying mechanisms might act through VL and VIM as the crucial relay stations.

Methods

We selected 20 patients with Parkinson's disease (PD) who gave their informed consent for thalamotomy. The procedures were previously described,[3] and were approved by the review board of the Neurological Clinic in accordance with the principles for research involving humans in the Declaration of Helsinki.

Tables 1 and S1 list the basic information of patients and surgery including the maximum estimates for rigidity, tremor, and bradykinesia before and after the surgery with the motor subscale in the unified idiopathic Parkinson's disease rating scale (UPDRS),[4] partially referring to the revision of original scale.[5] The point reductions occurred more for tremor (t-test, P < 0.0001) and rigidity (P < 0.0002), and less for bradykinesia (P < 0.002). Table S1 includes the side, number of sites and coagulation time of surgical lesion with the intercommissural distance (ICD), the half-width of third ventricle (HWV), and the lesion sites on the Schaltenbrand–Bailey atlas.[6] The lesion covered VL and VIM, as shown in the preceding papers.[1-3]

Table 1. Status of tremor, rigidity and thalamic activities in patients with Parkinson's disease
Patient ID no.Tremor, pre/postVτ (%)Cycle-τ/tOn-τ/tCgw R, pre/postCgw-τRigidity, pre/postVβ (%)β/EMGStretchVoluntary contraction
βoffRelapsed ββoffRelapsed β
  1. +, Observed; –, not observed; +/–, preoperative presence/postoperative absence; β/EMG: coincidence of β-band activities and tonic electromyograms at rest (+), contingent synchronisms of β-band activities leading to discrete spikes of electromyograms in cyclic sequence (++), and additionally with positive correlation of β-waves with compound spikes of electromyograms in amplitude (+++); βoff, block of pre-existing β-band activities on muscle stretch or voluntary contraction; Cgw-τ, τ-range activities accompanying cogwheel rigidity; Cgw R, cogwheel rigidity; Cycle-τ/t, tremor lags behind τ-range activities in each cycle of recurrence; nt, not tested under suitable conditions; On-τ/t, tremor lags behind τ-range activities in initiation; pre/post: preoperative/postoperative unified idiopathic Parkinson's disease rating scale; Relapsed β, late rebound of β-band hyperactivities; Vβ, average rate of β-waves; Vτ, maximum rate of τ-range wave.

PD0853/0100+++/–+2/092.8++ntnt
PD1191/0100+++/–+2/069.4++ntnt
PD1363/2100+++/–+2/0.578.7++++ntnt
PD0823/091.0+++/–+2/068.1+ntnt
PD0403/090.4+++/–+2.5/074.3+++++++
PD1542/087.9+++/–+0.5/052.6+++ntnt
PD1502/087.3+++/–0.5/065.7++++ntnt
PD0732/082.8+++/–+3/178.7+++ntnt
PD1382/072.6+++/–+1/091.0++++++
PD1111/029.3+/–+3/183.6+++++nt
PD0690/018.2+/–+3/078.6++++++
PD1330/012.7+/–+2/071.1++++ntnt
PD1351/05.5+/–+2/074.8++++++
PD0700/04.7+/–+3/0.573.6++++++
PD0571/00.0+/–+4/160.5++++ntnt
PD0660/00.0+/–+2/177.7++++ntnt
PD0721/10.0+/–2/0.575.0++++++
PD1260/00.0+/–2/086.0+++++nt
PD1410/00.0+/–3/173.3++++++
PD1610/00.0+/–2/078.5+++ntnt
Mean1.25/0.1544.12    2.18/0.3375.20     
SD1.16/0.4943.77    0.86/0.449.53     

Throughout the surgery under local anesthesia, the patient was alert enough to maintain conversations with the medical staff. Brain activities recorded by the bipolar concentric semi-microelectrode were divided into the local field potentials (LFP) and multiple unit spikes (MUS). The electrode was positioned so as to record the 3 to 7-Hz (τ-range) or 13 to 27-Hz (β-band) activities corresponding to any of the bilateral neck and contralateral limb muscles in tremor and in rigidity, respectively, shown on the surface electromyograms (EMG).

Table 1 lists the maximum amount (Vτ) of τ-range LFP (waves) and the mean amount (Vβ) of β-band LFP (β-waves) representative of each case, as rated with the time integrals of 3 to 7 and 13 to 27-Hz wavelets, respectively, from the 3-s records in the ventral thalamus.[3] Table 1 aligns the cases in decreasing order of Vτ.

Results

Resting, postural and kinetic tremor related with τ-range activities

The patient's postural or active movements modify tremor. We assume the existence of some peripheral and central mechanisms that control the development and maintenance of tremor with either results of well fit performance or vicious conflict between voluntary and involuntary movements. We are interested in the latter case as a clue to analyze the developing and maintaining mechanisms of tremor.

Figure 1a shows a record from a case (PD073) with limb EMG (traces 1–3), LFP in VIM (4) and MUS (5). The record starts with resting tremor limited in the wrist extensors (Fig. 1a, trace 2, shown by the bar i), proceeds with tremor in response to a slight postural change in the upper limb (bar ii) and develops to the episodes of kinetic tremor with burst EMG (bars iii and iv) under the repeated instructions for the voluntary upper limb contraction. The kinetic tremor there spreads over the arm and wrist muscles almost in synchrony (Fig. 1a, bars iii, iv; traces 1, 2). No tremor occurs in the lower limb (Fig. 1a, 3).

Figure 1.

Resting, postural and kinetic tremor preceded by τ-range activities. (a) Records are taken from PD073 with electromyograms and brain activities at the calculated nucleus ventralis intermedius site on the Schaltenbrand–Bailey map of the posterior (5.0), horizontal (1.4), and lateral measures in mm (9.4). Traces of electromyograms: brachial biceps (1), wrist extensors (2) and gastrocnemius (3) with a scale of mV. Traces of nucleus ventralis intermedius activities: local field potentials (4) and multiple unit spikes (5), respectively, with 50- and 100-μV calibrations. Time is given in trace 6 with a 100-ms scale. (b) Bars i–iv indicate pieces of trace to be superimposed. Traces i–iv aligned at the selected peaks of electromyograms as indicated by dots that correspond to those in (a). (b) Calibrations: mV scale for traces 1–3, 25 and 50 μV for 4 and 5 with an expanded time scale of 100 ms.

The three modes of tremor, resting, postural and kinetic, are synchronized with τ-range activities (Fig. 1a, traces 4, 5). For close inspection, we align four specimen sets of trace in Figure 1b (i–iv) by moving superposition on the line at the selected five peaks of muscle spike (Fig. 1a, dots). The muscle spikes in resting tremor at 5.0 Hz (Fig. 1bi, 2) lag behind the τ-range LFP or negative-going depolarizing waves (4) with negativity-dominant MUS (5). Postural tremor shows a similar set of EMG, τ-range waves and MUS, but with faster rhythms at 7.6 Hz (Fig. 1bii). The kinetic tremor is further intensified in high amplitude at increased 8.1–8.3 Hz involving all the upper arm muscles (Fig. 1biii, iv; 1, 2), being preceded by further exaggerated τ-range activities (4, 5).

As the temporal order in recurrence of the τ-range depolarization, concurrent MUS and tremor EMG was held in general, the central origin hypothesis[1] could be extended to all the modes of tremor. Additionally important is the plausibility that the thalamus with exaggerated τ-range activities is not only the key locus of tremor generation, but also acts as the relay station of peripheral and central control on the development and maintenance of tremor.

Variation of τ-range activities and tremor in maintenance

When the τ-range activities and tremor are maintained with varying amplitudes, we can examine their quantitative relationship.

Inspecting first a minor variability, we pick up three pieces of record in VIM (Fig. 2a–c, PD082), where the configuration of τ-range waves varies with concurrent MUS (3) and brachial biceps EMG (1). The records are superimposed in Figure 2i–iii at the negative troughs of t-range waves (dots in a–c and i-iii). The t-range waves (Fig. 2i–iii, traces 2) show there the concave (i), linear (ii) or convex slope (iii) of τ-range waves down to the trough of negativity (traces 2), yielding spikes of different amplitude on the way (traces 3). The configuration of τ-range waves is characteristically deviated from the sinusoidal form, perhaps reflecting the process of membrane depolarization under the balance of excitatory and inhibitory synaptic input volleys or the cationic currents arising through the membrane channels of thalamic neurons.

Figure 2.

Tremor maintained on variable thalamic τ-range activities. (a–c) Records from PD082 with electromyograms of brachial biceps (1), local field potentials in nucleus ventralis intermedius (2) and multiple unit spikes (3) at the posterior (5.0), horizontal (1.2) and lateral measures (9.7). Traces in (a), (b) and (c) are superimposed in (i), (ii) and (iii), respectively, on the line at the negative troughs of t-range waves indicated by dots in (a–c) and (i–iii). Noticeably, the depolarizing t-range waves (2) induce multiple unit spikes (3) in keeping a quantitative relationship with tremor electromyograms (1). Calibrations: 0.5 mV, 25 and 50 mV for traces 1, 2 and 3, respectively. A time scale of 100 ms in (c) for (a–c), and that in (i) for (i–iii), respectively.

These τ-range activities lead EMG (Fig. 2i–iii). Relatively small thalamic unit spikes in Figure 2i (3) fire in bursts on the way of τ-range depolarization (2), and synchronize with small spikes of EMG from the brachial biceps muscles (1). Large thalamic unit spikes (Fig. 2ii, 3) then newly participate in the coupling of τ-range depolarization (2) and relatively large spikes of EMG (1). As tremor becomes small in Figure 2iii, the large unit spikes remain there, but depart free from the coupling with the tremor cycle.

Other than the thalamic unit spikes with tremor rhythm, there are quite a few non-relevant units. We naturally aim at the confined lesion to the sites with highly rated tremor-locked MUS. However, even when some MUS fire free from the tremor cycle, they contingently join in the cycle. The plausibility that most neurons in the thalamic target participate in tremor generation has thus provided a critical rationale of thalamotomy (see 'Discussion').

Next, we followed the developmental course of tremor on a long sample record (PD085) shown in Figure 3a. We measured there the amplitudes and the instantaneous frequencies of τ-range waves (Fig. 3a, trace 4) and of paired compound spikes in the upper limb EMG (1, 2). As plotted in Figure 3, their amplitudes (Fig. 3b) and frequencies (Fig. 3c) were variable, but they increased together in each other. The wrist extensors and brachial biceps EMG were quantitatively related with τ-range waves when paired at the time of simultaneous occurrence (Fig. 3d). Their relationships were significant (see legends of Fig. 3), and were of sigmoid mode, as shown by the curves obtained by the locally weighted scatterplot smoother method (tension = 66).[7]

Figure 3.

Tremor electromyograms developing in amplitude and frequency with τ-range waves. (a) Record from PD085 with electromyograms: brachial biceps (1), wrist extensors (2), gastrocnemius (3) and nucleus ventralis intermedius: local field potentials (4) and multiple unit spikes (5) at the posterior (4.5), horizontal (4.0), and lateral measures (11.7), and time scale of seconds (6). Voltage calibrations are given for electromyograms, local field potentials and multiple unit spikes. Onset of paired τ-range waves and multiple unit spikes (double-head arrow) is indicated as 0, with a vertical line. (b) The amplitudes of tremor electromyograms and τ-range waves in progression for approximately 18 s. Correlation coefficients with time: 0.82 for extensors (n = 90); 0.91 for biceps (n = 90); 0.68 for τ-range waves (n = 93; P < 0.0001 for all). (c) The frequencies of tremor electromyograms and τ-range waves in progression. Correlation coefficients with time: 0.41 for extensors (P < 0.0001); 0.44 for biceps (P < 0.0001); 0.22 for τ-range waves (P < 0.05). Note that the pulsation artifact in trace 4 of (a) limits accurate estimates of local field potentials. (d) Amplitude relationship of electromyograms with τ-range waves. (e) Similar onset-shifted relationship. Shifted onsets of electromyograms are shown with two short arrows in (a). Correlation coefficients with τ-range waves: (d) 0.64 and (e) 0.62 for wrist extensors (P < 0.0001); (d) 0.58 and (e) 0.61 for brachial biceps (P < 0.0001).

As the τ-range activities start (indicated with 0 as the onset in Fig. 3a) three cycles ahead of tremor EMG, we assume a role of some accumulating process in building up tremor. Figure 3e plots the relationship of EMG to τ-range waves at their onset with the pairs shifted by three cycles. The relationships obtained there were similar to those in Figure 3d. These quantitative findings partly support the hypothesis that some intensifying mechanisms act to develop and maintain tremor through the thalamic τ-range activities.

Peripheral modulation on tremor through τ-range activities

The examiner's maneuver of tapping, pressing or stretching muscles is a part of routine examinations. The maneuver of holding the patient's limb often stops the pre-existing moderate resting tremor and τ-range activities in VIM. In this condition, the trial of stretch induces accelerated muscle tone loaded with tremor-like rhythms; that is, the cogwheel rigidity.

Figure 4a shows EMG (1–4), LFP (5), and MUS in VIM (6) on the alternatively applied phasic gentle stretches of wrist flexors (2) and extensors (3), as indicated by arrows (PD119). Three trials of stretch to flexors induced three bursts of compound spike in EMG (Fig. 4a, trace 2), being preceded by a large surge of negative LFP (5) with three bursts MUS (6). The superimposed traces in Figure 4b show more closely the pattern of responses of VIM (traces 5, 6) leading to the EMG burst (2) with a considerable time lag. We see here a gradually decaying phase of three or more bursts in coupled MUS (Fig. 4b, 6) and EMG (2).

Figure 4.

Cogwheel rigidity and thalamic activities. (a,c) Records of muscle stretch from PD119 with electromyograms: brachial biceps (1), wrist flexors (2), wrist extensors (3), gastrocnemius (4) and nucleus ventralis intermedius: local field potentials (5) and multiple unit spikes (6) at the posterior (3.7), horizontal (3.9), and lateral measures (11.1), and time scale of 100 ms (7). (b) Three trials of stretch to wrist flexors from (a) are superimposed with an expanded time scale. (c–e) Responses to extensor stretch with cogwheel rigidity (c) are superimposed at selected peaks of (d) local field potentials and of (e) electromyograms. Voltage calibrations are given for electromyograms, local field potentials and multiple unit spikes.

Meanwhile, two trials of stretch to wrist extensors induced irregular EMG burst (Fig. 4a, trace 3) led by a small negative shift in LFP (5) with little change in MUS (6). Another trial in Figure 4c of stretch with holding to wrist extensors induced a relatively long-lasting repetitive EMG burst (trace 3) as a typical pattern of cogwheel rigidity. These EMG were led by LFP (Figure 4c, trace 5) and MUS (6). Being superimposed at the negative peaks of LFP (Fig. 4d, dots in Fig. 4c) or at the spikes in EMG (Fig. 4e, dots in Fig. 4c), the trial shows the correspondence of LFP to EMG with some time lag in between (filled to open circles in Fig. 4d, or open to filled circles in Fig. 4e).

The VIM-leading slight or moderate cogwheel rigidity shown here suggests the existence of some cumulative processes in the central circuits including the thalamus as a relay station to drive the tremor-like EMG bursts as the final outcome. By showing the depolarizing LFP and MUSs the records in Figure 4 prove that the kinesthetic afferents from muscles deliver excitatory synaptic inputs to the VIM neurons.

Stretch-induced cogwheel and tonic rigidity with τ-range and β-band activities

The cogwheel rigidity accompanies exaggerated muscle tone. Figure 5 shows close correlations of VIM τ-range activities with cogwheel rhythms in EMG and of VL β-band activities with rigid tone in EMG during a course of electrode navigation from VL to VIM regions (PD135).

Figure 5.

Stretch-induced cogwheel and tonic rigidity with τ-range and β-band activities. (a,b,d,e) Records are taken from PD135 with brachial biceps electromyograms (1) and nucleus ventralis intermedius local field potentials (2) and multiple unit spikes (3) at positions, respectively, shown with numerals in mm above the intercommissural line along the track and with calculated axes at (a) posterior measurement (Pm; 3.1), horizontal measurement (Hm; 4.6), lateral measurement (Lm; 11.1); (b) Pm 3.5, Hm 4.1, Lm 11.0; (d) Pm 4.0, Hm 3.4, Lm 10.9; and (e) Pm 5.8, Hm 1.1, Lm 10.6, and time scale of 100 ms (4). (c) Superimposed records from (b; bar iii and dots) to show synchronism of electromyograms to β-band activities. Vertical arrows in (a, b, d, e) indicate the onsets of muscle stretch. (b,d) Bars i, ii and iii show three sequential phases of response with changes in β-band activities. (f) superimposed records from (e; dots) show synchronism of electromyograms with τ-range activities. Voltage calibrations are given for electromyograms, local field potentials and multiple unit spikes.

When the electrode entered in VL (5.0 mm from the prescribed point on the intercommissural line), small β-band activities (Fig. 5a, traces 2, 3) appeared with little change on the two trials of muscle stretch indicated with two vertical arrows. These trials in the brachial biceps (Fig. 5a, 1) first induced the early tremor-like cogwheel and late tonic β-band rhythmic EMG, and then induced the continuing cogwheel EMG.

In the midst of VL (Fig. 5b, 4.26 mm), large β-band activities occurred in two bursts (2, 3) to muscle stretch, as indicated by bars ii and iii. Superimposed traces in Figure 5c show the second burst of β-band activities (traces 2, 3; marked with dots in Fig. 5b) and synchronized EMG (1). Further lowering the electrode (Fig. 5d, 3.27 mm), we saw the VL activities (2, 3) responding to muscle stretch (arrow in 1) with three steps: first suppression of pre-existing β-band activities (bar i) followed by second (bar ii) and third (bar iii) relapses of β-band activities. Notably, the muscle stretch in Figure 5d caused cogwheel EMG (1), but with no corresponding τ-range activities in VL. These cogwheel EMG appeared to mask some part of the rigid β-band EMG led by VL β-band activities.

When the electrode reached VIM (Fig. 5e, 0.0 mm), we saw τ-range activities (2, 3) occurring in response to muscle stretch, as shown in the preceding Figure 4. These VIM activities lead to cogwheel EMG (1), as shown in Figure 5f with the traces superimposed at the selected peaks of EMG (dots in Fig. 5e,f).

The sequential pattern of phasic suppression and tonic relapse of β-band activities in Figure 5b and d characterizes rigidity becoming manifest by muscle stretch. It implies that muscle stretch acts on rigidity in both ways of the off and on control. The on control acts longer to maintain tonic rigidity. We have often observed that the trials to push, stretch and tap muscles augment VL β-band activities to lead accelerated tonic rigidity without a cogwheel. The maintaining mechanisms of rigidity thus include the spinal servo apparatus that transmits kinesthetic afferent inflow to the thalamus as a key relay station.

Centrally modulated β-band activities and rigidity

In Figure 6a, a patient (PD040) is instructed to contract their brachial triceps (2, time of instruction indicated with a vertical arrow) with resultant twitches more or less spreading on the upper limb muscles (1–4).

Figure 6.

Phasic and tonic influences of active muscle contraction on β-band activities and rigidity. (a) Records were taken from PD040 with electromyograms of brachial biceps (1), brachial triceps (2), wrist flexors (3), wrist extensors (4), anterior tibial muscles (5), and local field potentials (6) and multiple unit spikes (7) at the posterior(2.7), horizontal (3.8), and lateral measures (9.0). Nucleus ventralis lateralis β-band activities before and after active brachial triceps muscle contraction (instructed at the time shown by a vertical arrow) show sequential suppression (bars ii to i) and relapse (iii). The parts of (a) ii and iii are superimposed in (b) and (c), respectively, at the selected peaks of local field potentials shown by dots in (a). Voltage and time calibrations are given for electromyograms, local field potentials and multiple unit spikes.

When the patient sets in motion, VL activities show the pre-existing β-band activities wiped off and the negative-going LFP with fast discharges of MUS prompted instead (Fig. 6a, traces 6, 7), forming a preparatory phase for the later appearance of voluntary phasic EMG (2). With these changes (Fig. 6a, bar i), active motor initiatives lead to voluntary contraction and counteract rigidity.

The active suppression of rigidity is then followed by vigorous relapse of β-band activities (Fig. 6a, 6, 7) leading to long-lasting recurrence of rigid EMG (1, 2, 4). The superimposed traces in Figure 6b and c show contrast of the β-band activities before (Fig. 6b, indicated with bar ii and dots in Fig. 6a) and after the voluntary contraction (Fig. 6c, with bar iii and dots in Fig. 6a). The EMG of traces 1, 2 and 4 in both Figure 6b and c are all synchronized with β-band activities (6, 7), and the brachial triceps (2) among them appears to reflect the change in strength of β-band activities higher in amplitude of LFP (6), and with more MUS (7) in Figure 6c than in Figure 6b.

When the instruction urges the patient's continuous effort to move and hold, strong contraction occurs for a long time. Figure 7a shows the trials of wrist flexion and extension in sequence (two vertical arrows). The first flexor trial results in fairly good tonic contraction of flexors (trace 4), but with co-contraction of the antagonist extensors (5). The patient's maximum effort is manifest in initially bursting muscle contractions that spread over evenly to the neck muscles (Fig. 7a, ipsilateral sternocleidomastoideus in trace 1).

Figure 7.

Voluntary and involuntary movements in conflict as shown by thalamic β-band activities repeatedly relapsing. (a) Records are taken from PD040 with electromyograms of the sternocleidomastoideus (1), brachial biceps (2), brachial triceps (3), wrist flexors (4), wrist extensors (5), anterior tibial muscles (6), and local field potentials (7) and multiple unit spikes (8) at the posterior (2.7), horizontal (3.8), and lateral measures (9.0). (a) Successive trials of the patient's maximum efforts to flex and hold his wrist (left arrow) and to extend and hold his wrist (right arrow). The first trial resulted in tonic contraction of flexors (trace 4), but with co-contraction of antagonist extensors (5). The second trial resulted in failure to keep tonic contraction of extensors (5) as a result of repeated relapses of β-band activities (7, 8) and thereby disrupted electromyograms (5; interrupted line), but without antagonistic contraction of flexors (4). (b–e) Superimposed traces in parts indicated, respectively, by assemblies of dots ii–v in (a). Variable β-band activities (7, 8) synchronize there with upper limb electromyograms in different degrees (1–5). Voltage and time calibrations are given for electromyograms, local field potentials and multiple unit spikes.

Responding to the instruction, VL activities show suppression of pre-existing β-band activities, and bursting of motor initiative MUS (Fig. 7a, bar i; 7, 8) followed by phasic and tonic voluntary wrist contractions (4, 5) with minor interruption as a result of temporarily recurring β-band activities (7, 8).

The second extensor trial (Fig. 7a) shows a similar sequence of phasic and tonic voluntary contraction of wrist extensors (5), but often interrupted, masked or weakened in its tonic phase of EMG (interrupted line) phase-locked with vigorously repeated surge of β-band activities (7, 8). The movement itself faithfully followed the instruction without antagonistic contraction of flexors (Fig. 7a, 4). The episode of failure in keeping long-lasting contraction of extensors shows voluntary and involuntary movements in conflict, which is caused by counteraction between the motor initiative activities and excess β-band activities in VL.

Several pieces are selected from the traces in Figure 7a to examine the synchronism of β-band activities with EMG, as shown by the dot assemblies of ii–v. Figure 7b–e shows superimposed traces, where variable β-band activities (7, 8) synchronize with the upper limb EMG in different degrees (1–5), and most clearly with the main agonists in quantity (4, 5) that are instructed.

Viewing through the records of MUS in Figures 6 and 7, it appears that the spikes set for voluntary movements and those of relapsed β-band activities inducing involuntary disturbance are different in amplitude. These dual changes of VL activities would be brought about by differentially physiological and pathological groups of VL neurons.

Peripheral and central actions on tremor and rigidity summarized

In Table 1, tremor, rigidity and thalamic activities are summarized in their status at rest or modulated on the list of 20 cases aligned in the decreasing order of Vτ.

The preoperative score of tremor on UPDRS is correlated with Vτ (correlation coefficient, 0.865, r < 0.0001). The cases with moderate tremor (Table 1, PD085, 119, 136, 082, 040, 154, 150, 073, 138) show the τ-range activities leading to tremor recurrence (Cycle-τ/t+) and initiation (On-τ/t+). When postural and kinetic modes of tremor occur, their frequencies are increased compared with the resting mode. All the cases show cogwheel rigidity (Cog R+), mostly with stretch-induced τ-range activities in VIM (Cgw-τ+).

In all the cases with rigidity in Table 1, the rated VL β-band activities (Vβ) are as high as 60% or more, in phase-related with tonic EMG (β/EMG+), mostly synchronized with EMG (β/EMG++) and in a moderate quantitative correlation with EMG in two cases (β/EMG+++). Both passive and active muscle manipulations caused transient suppression (βoff), and later a relapse of β-band activities (Relapsed β) in most cases when tested at the recording sites that somatotopically met the muscles.

Discussion

Thalamotomy is effective in treating parkinsonian tremor and rigidity. Yet, the rationale has not been established on its invasive means in that the surgery aims to remove neurons located far away from the brain regions directly affected by degeneration. The symptoms should be explained by their generation and maintenance by referring to how the surgical targets are involved.

We have tested several relevant lines of the working hypothesis as follows: (i) neurons driving tremor and rigidity are localized in the ventral thalamus including VIM and VL, and the lesion there stops tremor and rigidity; (ii) many nearby thalamic neurons fire normally in conditions of active and passive motor control, but most of them contingently become highly active to lead to tremor or rigidity; (iii) tremor and rigidity are both generated in the central neural circuits including the thalamus; and (iv) tremor and rigidity are maintained peripherally as well as centrally through the thalamus.

For these hypotheses, the previous[1-3, 8] and present analyses have accumulated partial and circumstantial evidence. Neurons driving tremor have been the long-run hallmark of thalamotomy. We now know the neurons driving rigidity that exaggerate β-band rhythmic discharges in neck and limb EMG.

The thesis of contingency (ii) maintains our experience. We see some VIM neurons firing out of the tremor cycle and some VL neurons firing in, preparing the instructed voluntary contraction. The possibility of removing some healthy neuronal population has urged us to minimize the lesion. Actually, the minimal lesions, such as mechanical damage caused with the electrode, are too small to affect the whole figure of symptoms.[1] For the benefit of individual cases, we have compromised the amount of lesions. As summarized in Table S1, the number of lesion sites and the coagulation time required for remedy are thus widely varied.

The generating mechanisms of tremor and rigidity with their characteristic rhythms have been unknown, despite some hypotheses proposed so far.[9, 10] The thalamic nuclei, VIM and VL might be involved as either a site of genesis or a station of output to the upstream brain regions, such as the cerebellum and basal ganglia.[9-12] Perhaps, more important is that these nuclei are the crucial zone for the maintenance of tremor and rigidity to which both the central and peripheral intensifying inputs converge. The cogwheel rigidity, in particular, relies on both VIM and VL through their marker activities.[13] By setting VIM and VL on the thalamocortical route projecting finally to the spinal cord, our findings partly provide evidence to support the earlier proposed hypothesis of hyperactive transcortical long-loop reflexes contributing to tremor and rigidity.[14-17]

Acknowledgments

The materials of this study have been supplied from the collective works of the late Dr Hirotaro Narabayashi. The authors thank Ms Sachiko Nakata for her assistance in managing the clinical records of Narabayashi Memorial Laboratory of Neurology.

The authors declare no conflicts of interest.

Ancillary