Elizabeth S. Lorenc1*, Kartik Sreenivasan2, Derek E. Nee3, Annelinde R. E. Vandenbroucke4, & Mark D’Esposito1,5
1Helen Wills Neuroscience Institute, University of California, Berkeley, California, USA
 
* Correspondence: Elizabeth S. Lorenc, Helen Wills Neuroscience Institute, University of California, Berkeley, 132 Barker Hall, Berkeley, CA, 94720-3190, USA.
elizabeth.lorenc@berkeley.edu
Keywords:
 
Abstract
 
 
 
 
 
 
 
 
 
 
1.              Introduction
Visual working memory (VWM) allows for the maintenance of precise visual details of objects no longer in view, and is supported by activity in many brain regions, including lateral prefrontal cortex (D’Esposito et al., 1995; D'Esposito, Postle, & Rypma, 2000) and primary sensory cortices (Harrison & Tong, 2009; Serences 2009).
While “sensory recruitment” models (D'Esposito, 2007; Postle, 2006) suggest that precise visual details are maintained in stimulus-selective primary visual regions, it remains unclear what happens to VWM representations in the face of subsequent visual input.
To this end, the present experiment
 
2.              Materials and Methods


2.1.         Participants


__ participants completed the training, ___ participants completed a retinotopic mapping session, and ultimately a total of twelve healthy young adults (mean age___, 2 male) completed the entire experiment. Each of the twelve participants completed a one-hour training session and four two-hour MRI scan sessions. All procedures were approved by the UC Berkeley Committee for the Protection of Human Subjects. Participants gave their written informed consent before the study and were compensated monetarily for their time.
 
2.2.         Cognitive task
 
 
Timing (slightly different timing for first two participants), method-of-adjustment response, feedback at the end of each run, pay for precision.
 
Orientations covered the whole 180 degree space – 8 base orientations, with +/- 1 - 10 degrees of jitter: ensured that the same general set of orientations were shown in each of the three distractor conditions. Broken up between runs, so that half of the orientation categories were shown on odd runs and half on even runs. 12 trials per run: one of each of 4 orientation categories and 3 distractor conditions.
 
One third of trials did not have a distractor stimulus, and the rest had a distractor with an orientation that was 40 – 50 degrees clockwise (50%) or counterclockwise (50%) of the remembered orientation. With the exception of two of the authors, participants were unaware of the relationship between the memory and distractor stimulus orientations.
 
parameters of the gabors: size, eccentricity, full contrast, degrees of visual angle, phase-alternating

Figure 1. Right-lateralized delayed-estimation task for orientation, with intervening distractors
 
Stimuli were projected onto a screen at the rear of the magnet bore and viewed via a headcoil-mounted mirror, and responses were collected with an MR-compatible joystick (Current Designs, Inc.).
 
Participants were required to maintain central fixation throughout each scanning run, and eye position was continuously monitored using an MR-compatible eyetracker (Avotec).
 
2.3.         Behavioral analyses
 
The method-of-adjustment response yielded a trial-by-trial measure of memory error in degrees for each distractor condition (no distractor, clockwise distractor, counterclockwise distractor), for each participant. Using the MemToolbox (citation) in MATLAB (Mathworks), each error distribution was fit with a mixture model of a von Mises distribution and a uniform distribution. As has been previously published (Zhang & Luck, 2006), three free parameters were estimated: the mean of the von Mises (reflecting any systematic clockwise or counterclockwise biases in participants’ responses), the standard deviation of the von Mises (reflecting the average precision of a participant’s responses), and the height of the uniform distribution (reflecting the rate of random guesses). This model was fit separately for each distractor condition, but hierarchically across all twelve participants, which in addition to yielding participant-specific parameter estimates, also
 

2.4.         Functional MRI acquisition and preprocessing


MRI data were acquired in the UC Berkeley Henry H. Wheeler, Jr. Brain Imaging Center with a Siemens TIM/Trio 3T MRI scanner with a 12-channel receive-only head coil. Whole-brain MP Flash T1-weighted scans were acquired for anatomical localization and normalization. Functional data were obtained using a one-shot T2*-weighted echoplanar imaging (EPI) sequence sensitive to blood oxygenation level-dependent (BOLD) contrast. The EPI sequence parameters for the first two participants (TR = 1.6667s, TE = 30ms, field of view = ??, matrix size = ??, in-plane resolution = 3 x 3mm, ?? contiguous 3mm-thick axial slices separated by a 0.3mm interslice gap) were slightly different from the sequence used for the remainder of the participants (TR = 2.0s, TE = 30ms, field of view = ??, matrix size = ??, in-plane resolution = 3 x 3mm, ?? contiguous 3mm-thick axial slices separated by a 0.3mm interslice gap), as adjustments were made to improve whole-brain coverage.
Gray/white matter boundary segmentation and cortical surface reconstruction was performed with Freesurfer’s (citation) recon-all tool, and all surface-based analyses were then performed in AFNI’s SUMA (citation) package.
Functional MRI data were then subject to standard preprocessing with AFNI (Cox, 1996; Cox & Hyde, 1996) and custom Matlab (v2011b, The MathWorks, Inc., Natick, MA) scripts. Motion correction and volume registration of each EPI run to the anatomical scan was carried out in a single resampling step by align_epi_anat.py (Saad et al., 2009), by first aligning the mean of the middle EPI to the anatomical data and then aligning each volume to that mean EPI with a 12-parameter affine registration. Finally, each run was z-scored temporally, voxel-wise, in preparation for forward encoding modeling (FEM) and multi-voxel pattern analysis (MVPA). 
 
2.5.         Region-of-interest creation
 
Spatial localizer & retinotopy stimuli
 
Left and right visual areas V1 – V3, dorsal and ventral, were delineated on the surface based on separate polar angle and eccentricity mapping scans, following standard procedures (cite @Retino_Proc).
2.6.         Forward encoding model analyses
 
The following analyses were performed separately for the left and right early visual areas (V1 – V3). Because the stimuli were always presented in the right hemifield, the left hemisphere was always contralateral to the memory stimulus, distractor, and probe, and the right hemisphere was always ipsilateral.
 
Extract two volumes representing stimulus perception (__ to __s after stimulus onset), the first memory delay (__ to __s after stimulus onset), distractor perception, and the second memory delay, and average each pair to yield a single BOLD intensity pattern for each trial epoch, for every trial. The forward modeling analysis was then completed separately within each trial epoch, using a leave-one-run-pair-out cross-validated structure.
 
Basis set: 8 channels, shape of channels,
 
How significance was evaluated:
 
Control: delta functions
 
 
 

2.7.         Multi-voxel pattern classification analyses
 
Regularized logistic regression classifier – penalty
 
Look at classifier evidence – provides a more nuanced measure than raw accuracy
 
 
 
3.              Results
 
3.1.         Behavior
 
 
3.2.         Orientation reconstruction in early visual areas
 
3.2.1.     Sinusoidal basis functions
Can reconstruct perception only in ipsilateral hemisphere (do stats to compare directly), but can reconstruct orientation bilaterally during the memory delay.
 
Effect of distractor on reconstructions
 
Any way to see if there is a bias??
 
 
3.2.2.     Delta basis functions
 
 
4.              Discussion
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5.         Acknowledgments
Funding:
 
6.         References
9.         Tables and figures
9.1    Figure legends