Pharmacological therapies currently marketed for Alzheimer’s disease (AD) are only symptomatic and show limited effects in terms of clinical benefit. Thus, the development of new symptomatic drugs remains essential. However the dramatic increase in costs associated with drug development together with the poor number of emerging drugs highlights how crucial it is to accelerate the findings aiming to bringing new drugs to market. In this respect, optimization of the development process by integrating, at early stage, reliable biomarkers able to predict clinical benefit in phase III clinical trials may help. The improvement of certain techniques such as imaging and electrophysiological methods has led to a more accurate assessment of the brain’s physiological impact of pharmacological treatments used to alleviate symptoms in AD patients. This review aims to gather the main findings from clinical studies where the effect of anti-dementia drugs were assessed in healthy volunteers and AD patients through one or several such biomarkers (electroencephalography (EEG), magnetic resonance imaging (MRI), positron emission tomography (PET), single-photon emission computed tomography (SPECT)). Overall, evidence presented in this review suggests that various biomarkers associated with key impairments observed in AD were sensitive to acetylcholinesterase inhibitors (AChE-I) medication and memantine with a good correlation with enhancement of cognitive performance. In most of the reviewed studies, only one kind of biomarker was used. Among these, deficits in quantitative EEG profile, P300 latency, and regional brain activity measured by either functional MRI (fMRI) during face encoding and working memory task or by PET/SPECT have been shown to be reversed by anti-dementia drugs. It is therefore suggested that a single biomarker approach would be limited and not be sufficiently predictive to extensively assess the potential of a new symptomatic drug. Hence, it appears that a combination approach with the use of a panel of biomarkers rather than a single biomarker may be more appropriate to establish a good correlation between the disease and therapeutic intervention.