It remains unclear whether distinct anatomical brain regions within the medial temporal lobe (MTL) subserve different functional roles during successful memory encoding and retrieval (1-5). In order to address this gap in current knowledge, we designed an fMRI experiment to investigate the neural basis of associative memory processes during face and name linking. Our approach was to compare and contrast successful-recall and failed–recall in order to delineate brain regions in the MTL that make distinct functional contributions during face-name associative memory formation.

Subjects: Twenty (10 females; age range= 21–45 yrs; mean age = 31yrs) participated in the study with informed consent.

fMRI Paradigm: All participants completed a block design, face-name learning and recognition fMRI paradigm. The face-name task included 16 faces (8 male) and 64 face-name pairs. During the 16 s encoding trials, a face-name pair was presented on a LCD screen. During the 16 s retrieving (or recognition) trials, a face was presented with 4 names on a LCD screen. A 48 s distractor period, alternating between “+” and “0” was included after each encoding trial (Figure 1). During recognition trials, participants were asked to indicate with a button press which name was associated with the face during the learning (or encoding) phase. Participants were presented with face-name pairs during each learning block and similarly, 4 face-name pairs during each recognition block.

Equipment and scanning parameters: fMR images were acquired with a 3T Siemens Trio and 8 channel head coil with a T2*-weighted gradient-echo EPI sequence with the following parameters: TR/TE/FA = 2000ms/30ms/90 deg; FOV = 240 mm× 240mm; Matrix size = 80 × 80; # of axial slices=35; slice thickness = 4 mm; # of repetitions = 174.

Data analysis: fMRI data were processed using SPM8 (Wellcome Trust Centre for Neuroimageing). Activation maps were generated using the General Linear Model (GLM). A block was deemed a success (S) if all four face names were recalled correctly. All other blocks were marked as failed-recall blocks (F). The distractor period was taken as the baseline block (B). Subject wise contrast maps were generated for each experimental condition, i.e., learning or recall, the S-block or the F-block in comparison to the B-block. Group-level maps were then generated using a second level random effects analysis.

Significantly higher activations were detected in the right posterior hippocampal formation (P < 0.0043), left anterior hippocampus (p < 0.0133), left amgydala (p < 0.0207), left perirhinal (p < 0.0272), right parahippocampus (p < 0.0066) during the S-Block when compared to the F-block (Figure 2). The fMRI % signal change in the left posterior hippocampus during the experiment showed a rapid decrease for each block of learing and recognition (Figure 3).Our results highlight that no medial temporal lobe (MTL) structure can independently process and perform the cross-domain association memory demands of face-name associations. This process seems to be highly dynamic when quantified in terms of the BOLD signal time-course in respective brain structures. Taken together, these results suggest an intricate interaction between MTL structures for learning and recognition memory manipulations in the human brain. In turn, this will help ask new questions about human memory formation that were not feasible before.