multimodel tms - finding biomarkers for epilepsy
Annika de Goede is a PhD student in the Clinical Neurophysiology Group. Her supervisor is prof.dr.ir. M.J.A.M. van Putten from the Faculty of Science and Technology.
Epilepsy is characterized by the occurrence of epileptic seizures, resulting from an imbalance between excitatory and inhibitory brain activity. Around 5% of the population experiences an unprovoked seizure during their lifetime. However, less than half of them actually develop epilepsy. Patients are diagnosed with epilepsy when they had two or more seizures or when they had one seizure with an increased risk of recurrence. After a first seizure it is therefore important to estimate the risk of more seizures. Finding epileptiform discharges in the electroencephalogram (EEG) reflects a tendency to generate seizures. However, if no abnormalities are found, uncertainty remains as the sensitivity of EEG recordings is limited. Indeed, in about 10% of patients with epilepsy, abnormalities are not found despite repeated or prolonged recordings. This motivated our search for a novel biomarker to improve the diagnostic process.
Transcranial magnetic stimulation (TMS) is a technique to assess the balance between excitation and inhibition. It uses magnetic pulses to stimulate the brain. Stimulation consists of giving one pulse at a time (single pulse TMS), or two pulses short after each other (paired pulse TMS). The effect of stimulation can be assessed indirectly using the muscle response (motor evoked potential: MEP), or directly using the brain response (TMS evoked potential: TEP). If the interval between paired pulses is 50-400 ms, the MEPs and TEPs are inhibited (long intracortical inhibition: LICI). In this thesis, all these modalities were combined. We applied single and paired pulse TMS and evaluated the MEP, TEP and LICI. Since epilepsy is associated with increased cortical excitability, we expect the muscle and brain responses to differ from those in people without epilepsy.
This thesis describes our first steps towards the implementation of a multimodal TMS approach in epilepsy. It focusses on the clinical feasibility of multimodal TMS (Chapters 2 to 4), on finding biological modulators of cortical excitability (Chapter 5) and on the diagnostic value of multimodal TMS in epilepsy (Chapters 6 to 8).
Chapter 2 evaluates the repeatability of paired pulse TMS, using the muscle response. We measured thirty healthy subjects twice, about a week apart. On a group level repeatability of LICI was good, whereas individual subjects showed a large variation in LICI repeatability. In approximately half of the subjects we measured similar LICI both times, while LICI was (completely) different one week later in the others. This makes it more difficult to follow patients over time using TMS, for example to evaluate the effect of anti-epileptic drugs.
Chapter 3 evaluates the effect of changes in coil positioning during single and paired pulse TMS, using the muscle and brain response. To give the magnetic pulses, a TMS coil is placed on the head of the subject. The easiest and fastest way to place and hold the coil in position is manually, with a risk that it moves during the session. On a group level, we found no significant effect of a 5 mm change in coil location on the MEP, TEP or LICI, nor of a 10° change in orientation on LICI. This indicates that manual coil positioning performed by an experienced investigator is sufficient in clinical practice.
Chapter 4 evaluates the repeatability of single and paired pulse TMS, using the brain response. Similar to the muscle response, repeatability of the TEP was good after one week. In addition, we found a similar distribution pattern for the characteristic components of the single and paired pulse TEP. A difference between both responses was strong inhibition of the late paired pulse TEP components. We used these findings in healthy subjects as reference values for our epilepsy patients.
Chapter 5 explores if infraslow activity modulates cortical excitability. Giving the same magnetic pulse multiple times, not necessarily results in equal MEPs and TEPs. Very slow biological fluctuations may cause responses to vary over time. This large response variability may limit the applicability of TMS. Our findings indicate that infraslow activity contributes to the variability, although other mechanisms are likely involved as well. By only giving pulses during specific periods of the infraslow activity, it might be possible to reduce the response variation.
Chapter 6 provides a systematic overview of the current single and paired pulse TMS findings in epilepsy patients. So far only muscle responses have been evaluated in epilepsy patients who were not (yet) taking any medication. Findings were most consistent for paired pulse TMS. Compared to healthy subjects, cortical excitability was significantly increased in epilepsy patients at the intervals 250 and 300 ms.
Chapter 7 evaluates differences in the muscle response between refractory epilepsy patients and healthy subjects. Around 30% of the epilepsy patients continues to have seizures, despite taking anti-epileptic drugs. Based on combined LICI data from four different centers, we could not differentiate refractory epilepsy patients from healthy subjects. We were unable to confirm previous findings of increased cortical excitability in epilepsy patients.
Chapter 8 explores the potential of multimodal TMS to improve the diagnostic process. We measured patients short after their first seizure. Some of them were eventually diagnosed with epilepsy and some not. We could differentiate patients diagnosed with epilepsy from those without epilepsy and from healthy subjects, using LICI. In contrast to previous findings, excitability was increased at an interval of 100 ms and decreased at 200 ms. For single pulse TMS, we found no significant differences between the three subject categories.
In conclusion, we showed the clinical feasibility and potential of multimodal TMS to improve the diagnostic process in epilepsy.