Dear Sirs,

The striking paradox of pure autosomal dominant hereditary spastic paraplegia (AD HSP), in contrast to capsular stroke or primary lateral sclerosis, for example, is that despite extensive corticospinal tract (CST) degeneration and prominent lower limb spasticity, leg weakness is not a prominent early feature of the disease. Therefore, other descending motor pathways presumably compensate for CST degeneration, though, until the recent article by Nonnekes et al. [1] in the Journal of Neurology, this assumption had remained unproven. Using startling acoustic stimuli (SAS), they showed that the reticulospinal tract (RST) is not only functioning in patients with HSP, but that it compensates for lower limb deficits of postural control caused by CST degeneration.

In most patients with pure AD HSP, electrophysiological evidence of CST disease is limited to the lumbosacral cord [2]. However, some AD HSP genotypes (e.g., SPG4) are associated with a more severe phenotype and have motor-evoked potential (MEP) abnormalities in the upper limbs [3–5], consistent with post-mortem evidence of CST degeneration at all levels, from the medulla to the lumbo-sacral cord [6].

We tested MEPs (see Supplementary Methods) in two patients (father and son) with SPG31 AD HSP (REEP1 exon 5 c.337C > T/p.Arg113X [7]). In patient 2 (age 42; disease duration 37 years) MEPs were ‘typical’ of pure AD HSP [8], with prolonged central motor conduction times (CMCTs) in the lower limbs (CMCT 22.2 ms) and normal CMCTs in the upper limbs. However, in patient 1 (age 68; disease duration 64 years) CMCTs were significantly prolonged in the upper limbs (Fig. 1).

Fig. 1
figure 1

Examples of rectified motor-evoked potentials (MEPs) obtained from a 68-year-old patient with SPG31/REEP1 HSP (black) and an age-matched control (grey), aligned to the stimulus. a Upper limb MEPs recorded from right first dorsal interosseous (R FDI) and extensor digitorum (R EDC) muscles and evoked with a single transcranial magnetic stimulus (indicated by an arrow) at 70 % of maximum stimulator output (MSO), equivalent to 1.9 times the active motor threshold (AMT; for details see Supplementary Methods). Each trace is an average of 20 individual MEPs. Voltage calibration bars apply to both patient and control MEPs. b Lower limb MEPs recorded from right extensor digitorum brevis (R EDB) and tibialis anterior (R TA) muscles using a stimulus strength of 1.6x AMT (70 % MSO). c Right upper limb (FDI) central motor conduction time (CMCT). d Right lower limb (EDB) CMCT. In c and d, peripheral motor conduction times were subtracted from the FDI and EDB MEPs shown in a and b, revealing the CMCT and the result plotted on a longer time base. Solid vertical lines show the mean CMCT within the normal population (aligned to the start), and the grey boxes between vertical dashed lines extend two standard deviations above and below the mean [14]. Arrow heads indicate MEP onset

Full size image

Given the absence of clinical upper limb weakness in patient 1, we measured EMG onset latencies in a Start-React paradigm to see whether the RST might be compensating for the CST deficit in the upper limbs [9–11] (Supplementary Methods), as illustrated in Fig. 2a. Experiments had the relevant institutional ethical approval and complied with the Declaration of Helsinki, and were performed in both patients and 11 controls, aged 56–82 years. Traditionally, the effects of SAS on the visual reaction time (VRT), and the visual start-react time (VSRT), are normalized, thus:

$$ \Updelta {\text{VRT}}\left( \% \right) = \frac{{\left( {\text{VRT - VSRT}} \right)}}{\text{VRT}} \times 100. $$
(1)
Fig. 2
figure 2

a Example of normal raw rectified EMG data recorded from right biceps brachii in a 61-year-old male control subject performing the Start-React task. The EMG burst from which the VRT was measured is plotted in black, that from which the ART was derived is plotted in grey and the VSRT response is plotted in red. b, c Mean ∆T SR /∆T AR ratios calculated from EMG onset latencies measured in Start-React experiments are plotted. b Right first dorsal interosseous (R FDI) and c right biceps brachii (R BB). Error bars are standard error of the mean. Note that in patient 1, the R BB ratio is significantly increased (one-sample t test; p < 0.01)

Full size image

The normal ∆VRT is ~50 % [9], and a ∆VRT of less than 50 % is indicative of disease affecting the RST. However, because we were interested in measuring any change in the gain of the RST output, accessed via auditory pathways, we have used a ratio that incorporates the auditory reaction time (ART) following a low-intensity sound, as follows:

$$ \Updelta T_{\text{SR}} /\Updelta T_{\text{AR}} = \frac{{ ( {\text{VRT }} - {\text{VSRT)}}}}{{ ( {\text{VRT}} - {\text{ART)}}}} $$
(2)

where ∆T SR is the shortening effect of a SAS on the visual reaction time and ∆T AR measures the shortening of reaction time provided by a non-startling auditory stimulus, which presumably does not activate RST pathways. The results of this analysis are shown in Fig. 2c, d. Patient 2, who had no evidence of cervical CST disease, had normal ∆T SR/∆T AR ratios. However, patient 1, who had MEP evidence of cervical CST disease, had significantly increased ratios, but only when measured from biceps brachii EMG, despite normal ARTs and VRTs (Supplementary Results/Fig. 2). This result supports the notion that the RST mitigates the effects of disease within the cervical CST. The RST appears to compensate by increasing its output gain by a factor of around 1.5. Although the RST does project to both proximal and distal upper limb muscles [12], the effects of SAS are only seen in distal muscles in some tasks [13], possibly explaining why we detected differences only in the biceps muscle. These observations suggest that therapeutic interventions aimed at increasing the gain of RST outputs could improve recovery from neurological disorders characterized by CST dysfunction.