Statistics from Altmetric.com
Reduction in the neuronal activity of the subthalamic nucleus leading to diminished excitation of the globus pallidum internum is associated with chorea-ballism in monkeys.1 Levodopa induced dyskinesias are currently thought to share a similar pathophysiology2 but recent findings also suggest that abnormal patterns of neuronal firing in the globus pallidum internum may be as relevant.3 Data from both parkinsonian monkeys and patients with Parkinson's disease submitted to lesion4 5 or functional blockade of the subthalamic nucleus6 are in keeping with such a general principle, but the threshold to induce dyskinesias in the parkinsonian state is higher than in intact animals.7 The case recently described by Figueiras-Mendez et al 8 is extremely interesting as it suggests that functional inhibition of the subthalamic nucleus by high frequency stimulation blockades levodopa induced dyskinesias. This is clearly at odds with the current pathophsyiological model of the basal ganglia.9 Thus, the finding of Figueiras-Mendez et al 8 rises the intriguing possibility that dyskinesias depend or are mediated by neuronal firing in a given region of the subthalamic nucleus, which was blocked by high frequency stimulation. Measurement of afferent synaptic activity by the technique of 2-deoxylucose (2-DG) uptake showed an increment in the subthalamic nucleus (compatible with increased inhibition from the globus pallidum externum), particularly in the ventromedial tip of the nucleus.9 This contrasts with the findings in monkeys with chorea induced by pharmacological blockade of the globus pallidum externum, in which 2-DG uptake was maximal in the dorsolateral portion of the subthalamic nucleus, where the sensorimotor region lies. A recent anatomical study10 also showed that the cortical-subthalamic nucleus connection is somatotopically segregated, so that fibres from the supplementary motor area project to the most medial portion and fibres from the primary and premotor areas terminate in the lateral region of the subthalamic nucleus.10 All this heterogenity may have pathophysiological relevance, one aspect of which could be the findings in the patient reported by Figueiras-Mendezet al.8 However, before the findings of this case may be used to sustain a new hypothesis on the role of the subthalamic nucleus in the origin of levodopa induced dyskinesias, there is a crucial issue to resolve—namely, the location of the tip of the stimulation electrodes.
There are several points leading us to question the actual site of action of the electrode: (1) Stimulation of the subthalamic nucleus in Parkinson's disease has been associated with the production of dyskinesias only relieved by reduction in levodopa intake.6 Moreover, Benabid et al who pioneered this technique, consider the induction of dyskinesias by high frequency stimulation of the subthalamic nucleus as a good indicator of a very positive response. (2) Fibres travelling to the thalamus from the globus pallidum internum are placed dorsocaudally to the subthalamic nucleus and could be blocked by high frequency stimulation. (3) When the recording electrode goes into a region caudally to the subthalamic nucleus in sagittal planes 11 mm or less, neuronal activity is characterised by action potentials of large amplitudes (0.5–1 mV) with low background activity, tonically firing neurons, and absent sensorimotor responses (“driving”). All these characteristics seemed to be present in the patient discussed here. Neuronal activity in the sensorimotor region of the subthalamic nucleus is different from the above but on occasions the distinction may not be easy.
Accordingly, it is very important to document in more detail the findings in the case of Figueiras-Mendez et al.8 Ideally we would like to see the trajectory and length of the different recording tracks, the effects of microstimulation, and the postsurgery MRI with measurements of the location of the tip of the electrodes. If, as assumed, the subthalamic nucleus was indeed correctly targeted in this patient, the pathophysiology of the basal ganglia will need to be revisited.
Figueiras-Méndez et al reply:
We thank Obeso et al for their comments regarding our recent report.1-1 In summary, they raised some interesting points which need further clarification.
Recognition of the electrical activity of the subthalamic nucleus was based on the following criteria: (a) high frequency discharge (25 Hz or higher) within the nucleus1-2 1-3; (b) a tonic (regular), phasic (irregular) or a rhythmic pattern of discharge1-2; (c) response to voluntary/passive movements.1-2 1-4 When rhythmic discharges were recorded irregular passive manipulations were performed or the patients asked to moved the limbs irregularly; (d) response to tremor activity. Positive cells were so considered based on the correlated activity with the EMG and the accelerometer recorded simultaneously. Artificial manual stopping by one experimenter (confirmed by visual inspection, silence in the EMG, and stoppage in the oscillating accelerometer) and/or spontaneous arrest in the tremor modified the firing frequency and discharge pattern or rhythmic cells corroborating the tremor nature of the cells; (e) the activity of the cells above the subthalamic nucleus in the thalamus and zona incerta with proper characteristics1-2; (f) a change in the background basal noise when entering the subthalamic nucleus. A higher activity is observed1-2; (g) the contrary is observed when leaving the nucleus. A lower background noise level; (h) the activity of substantia nigra pars reticulata cells when further lowering the microelectode. These cells discharge at high frequency at regular intervals as identified in patients1-2 and primates.1-5 All these points were fulfilled by the patient reported.
Considering the questions in the letter by Obeso et al, we make the following comments: (a) Action potentials of large amplitude are easily recognised from the rest of the recording cells, and are not very common. The recordings shown in the article have amplitudes less than 0.3 mV and could not be considered large amplitude potentials. We start to record activity from 3 mm before entering the subthalamic nucleus, traverse the length of the subthalamic nucleus, and go further down several mm to encounter substantia nigra pars reticulata cells. Changes in the background activity are clearly recognised and are higher when entering the subthalamic nucleus. Enough cells are recorded along the tracks experimented so as to recognise a large amplitude potential. The low background activity found in our recordings is only due to the better signal-to-noise ratio of the electrodes used. “Good recording electrodes” depend on many variables such as tip size, tip profile, insulation material, impedance, manufacture, etc.1-6 The signal-to-noise ratio of the cells in question has the same ratio as the subthalamic nucleus cell shown by Hutchinsonet al.1-2
(b) In our report, cells discharged tonically, but also other cells fired phasically, well differentiated by a profuse burst activity and identified by statistical means (autocorrelation and interval histograms).
(c) Motor responses and tremorgenic cells in line with the above mentioned criteria were found along the trajectory of the electrode. Unfortunately, this point was not mentioned in the paper. It would surely have changed the opinion of Obesoet al.
Considering the mentioned patient, a total of eight neurons were recognised as belonging to the subthalamic nucleus in the right hemisphere, with a mean frequency of 74 Hz (range 38–109 Hz). Four of them responded to passive and/or voluntary movements and one was considered tremorgenic. The stimulating electrode was placed in laterality 11. One track was performed. In the left hemisphere, two tracks were performed. One track was dismissed by the poor responding activity of the cells recorded. In the other track, nine neurons were recorded in the subthalamic nucleus (always following the above mentioned criteria) with a mean of 69 Hz (range 17–98 Hz). Five cells responded to passive and/or voluntary movements. One of them was also positive to tremor. The stimulating electrode was placed in laterality 12. The effect of the stimulating electrode is always tested in the surgery before cementing it and, only when the symptoms are considered of unquestionable benefit it is left in the chosen place. The final position of the electrodes, assessed by ventriculography, was as follows: (a) posteroanterior: 1.5 mm behind the mean point of intercommisural line, (b) height: 6–6.5 mm below the intercommissural line, and (c) lateral: 12 mm for the right hemisphere, and 11.5 mm for the left hemisphere.