OBJECTIVE: To identify the neural underpinnings of cognitive deficits associated with memory problems in amnestic mild cognitive impairment (aMCI). BACKGROUND: Functional magnetic resonance imaging (fMRI) is increasingly used to assess patients with aMCI and could potentially help predict conversion to Alzheimer disease, but imaging results so far have been inconsistent in identifying brain activation patterns in aMCI. There is an immediate need to identify the neural substrates of different memory components that are affected by aMCI. METHODS: We used fMRI to study 13 patients with aMCI and 15 healthy age-matched controls during an associative memory encoding and recognition task. The picture-pair memory task encompassed different types of recognition trials to investigate recollection versus familiarity, and manipulated the relationship between paired pictures to investigate semantic processing. RESULTS: Brain activation during both encoding and recognition was lower in patients than controls, with greatest implications in the medial temporal lobe during encoding. Patients also had much greater impairment of associative recollection than recognition based on familiarity, along with a failure to recruit regions that normally respond to violations of learned associations. Finally, patients' impaired semantic encoding was reflected by deficient activation of a left frontotemporal network responsible for elaborate semantic processes. CONCLUSIONS: We show that a simple fMRI task may be sensitive to deficits in different memory components in aMCI and could thus prove useful in the development of an fMRI tool to assess and monitor patients.
OBJECTIVE: To identify the neural underpinnings of cognitive deficits associated with memory problems in amnestic mild cognitive impairment (aMCI). BACKGROUND: Functional magnetic resonance imaging (fMRI) is increasingly used to assess patients with aMCI and could potentially help predict conversion to Alzheimer disease, but imaging results so far have been inconsistent in identifying brain activation patterns in aMCI. There is an immediate need to identify the neural substrates of different memory components that are affected by aMCI. METHODS: We used fMRI to study 13 patients with aMCI and 15 healthy age-matched controls during an associative memory encoding and recognition task. The picture-pair memory task encompassed different types of recognition trials to investigate recollection versus familiarity, and manipulated the relationship between paired pictures to investigate semantic processing. RESULTS: Brain activation during both encoding and recognition was lower in patients than controls, with greatest implications in the medial temporal lobe during encoding. Patients also had much greater impairment of associative recollection than recognition based on familiarity, along with a failure to recruit regions that normally respond to violations of learned associations. Finally, patients' impaired semantic encoding was reflected by deficient activation of a left frontotemporal network responsible for elaborate semantic processes. CONCLUSIONS: We show that a simple fMRI task may be sensitive to deficits in different memory components in aMCI and could thus prove useful in the development of an fMRI tool to assess and monitor patients.