TMS is a brief and intense magnetic field, created by a strong electric current circulating within a coil resting on the scalp; it penetrates human tissue painlessly and, if the current amplitude, duration and direction are appropriate, it induces in the brain (or in spinal roots, or in nerves) electric currents that can depolarize neurons or their axons. TMS can demonstrate causal relationships between the stimulated area and behaviour. It can be used to interfere with the neuronal activity underlying sensory processing, motor planning or execution, attention and visual processing or awareness, different forms of declarative and non-declarative memory, reasoning, mathematics, language and consciousness. This temporary disruptive effect of TMS on cortical function is often referred to as a ‘virtual lesion’ and is likely to occur because the stimulus transiently synchronizes the activity of a large proportion of neurons under the coil, thus inducing a long-lasting, generalized inhibitory postsynaptic potential that reduces cortical activity for the next 50 to 200 milliseconds, depending on stimulus intensity and duration. Moreover, repetitive TMS (rTMS) can produce changes – outlasting the period of stimulation – in cortical excitability⁹.

rTMS studies are increasing knowledge about correlation between anatomic structures and their functions. For example, they showed that dorsolateral prefrontal cortex (DLPCF) is implicated in working memory and in verbal episodic memory tasks.

Lesions of frontal lobes cause a reduction of learning and memory abilities but they do not cause the same memory loss as bilateral hippocampal lesions do. So, it has been demonstrated that prefrontal cortex (PFC) is involved in episodic memory tasks and that during encoding or retrieval, there are different areas of PFC that are activated with different patterns. It has been demonstrated that left PFC is involved in encoding processes; instead right PFC is involved in retrieval processes. This theory is known as hemispherical encoding retrieval asymmetry (HERA)^9-11. Nowadays, it is known that the HERA model is not so absolute but that there could be variations due to the nature of the presented stimuli, the strategies of memorization used and the grade of difficulty of the task. Recently, positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) studies have demonstrated that PFC activity tends to be less asymmetric in older than in younger adults (hemispheric asymmetry reduction in old adults, HAROLD). This change may help counteract age-related neurocognitive decline (compensation hypothesis) or it may reflect an age-related difficulty in recruiting specialized neural mechanisms (dedifferentiation hypothesis)^12,13. Functional neuroimaging techniques could not be used to study whether the loss of asymmetry during ageing reflects compensatory mechanisms or dedifferentiation processes; in fact, they show only the activation of a specific area during a task, but this activation could be related to the performance of the task without being the cause. Instead, this issue was addressed by using rTMS to assess causal relationships between performance and stimulated brain regions. Rossi et al. studied the effects of rTMS applied to the left or right DLPFC simultaneously with the presentation of memoranda on visuo-spatial recognition memory task, in a population of healthy subjects divided into two classes by age (<45 and >50 years). In the younger subjects, rTMS of the right DLPFC interfered with retrieval more than left DLPFC stimulation. The asymmetry of the effect progressively vanished with ageing, as indicated by bilateral interference effects on recognition performance. Conversely, the predominance of left DLPFC effect during encoding was not abolished in the older subjects, thereby indicating its causal role for encoding along the life span. Findings with rTMS suggest that the bilateral engagement of the DLPFC has a compensatory role on episodic memory performance across the age span¹³.

Physiological ageing is associated with more complex activations of the motor system to compensate for the reduced skill in motor performance^14,15. Therefore, TMS variables can provide useful clues to investigate causally subclinical changes of cortical excitability at the motor cortex level. Unfortunately, studies do not show converging findings, probably as a result of technical and experimental differences among them^13,16-18. Apart from assessing cortical excitability, TMS can be used to measure intracortical facilitatory and inhibitory mechanisms resulting from the balance among neurotransmitters, and dynamic synaptic adaptations to progressive neuronal loss. Motor threshold is generally reduced in Alzheimer’s disease compared to healthy age-matched controls and is similarly reduced in subcortical ischaemic vascular dementia^16,19,20, even if several reports have not found reduction of motor thresholds^21,22, with one study noting increased rather than decreased motor thresholds in Alzheimer’s disease²³. Global cortical hyperexcitability while the brain is ‘at rest’ in Alzheimer’s disease is a consistent result²², which seems to occur independently of GABAergic and cholinergic cortico-cortical mechanism dysfunction as reduced motor thresholds and intracortical inhibition (ICI) and short-latency afferent inhibition (SAI) are uncorrelated. Since TMS does not provide specific pathophysiological information but is sensitive to the ‘global weight’ of several neurotransmitters as well as subcortical and cortical motor inputs, interpretation of cortical hyperexcitability in Alzheimer’s disease is not yet possible.

The majority of these TMS results indicate that the cortex of Alzheimer’s disease patients is hyperexcitable, and that such hyperexcitability may offer clues for the differential diagnosis from other dementias in which the cholinergic deficit is not predominant. An intriguing issue that can be addressed by TMS is the study of cortical reorganization of the motor output in Alzheimer’s disease. Indeed, motor cortex physiological organization is early involved in Alzheimer’s disease, despite the lack of clinically evident motor deficits, which only appear later²⁴. This observation is not surprising considering that the motor cortex, through muscarinic receptors and cholinergic terminals, receives a major input from the nucleus basalis of Meynert²⁵. Mapping of motor output in mild Alzheimer’s disease patients asymptomatic for motor deficits has demonstrated frontal and medial shift of the cortical motor maps for hand muscles. In healthy controls the centre of gravity of motor cortical output is coincident with the site of maximal excitability (hot spot²⁶), whereas in Alzheimer’s disease patients it shifts towards frontal and medial regions in the absence of changes in the hot spot²². An altered synaptic plasticity in Alzheimer’s disease also has been demonstrated by applying brief trains of high-frequency rTMS to motor cortex and recording motor-evoked potentials (MEPs) from the contralateral hand muscles. This procedure can induce a progressive increase in MEP size in normal age-matched controls, but it produces opposite changes in Alzheimer’s disease. This finding was interpreted as an altered short-term synaptic enhancement in excitatory circuits of the motor cortex. Additional assessment of these findings will be required by evaluating Alzheimer’s disease patients after different pharmacological treatments (e.g. cholinergic versus drugs acting on NMDA glutamate receptors).