media.nature.com  · Web viewManual of Procedures Rationale The development of new intervention strategies in the prevention or delay of Alzheimer’s disease onset are an international

Supplementary Information

Method

MRI preprocessing methods

Hippocampal Manual Tracing: similar method details were presented elsewhere 1. To

minimize error, we firstly pitch-rotated a standard template so that the long axis of the

hippocampus was orientated horizontally on sagittal view. Each subject’s baseline and

follow-up T1 images were coregistered with each other, then coregistered to this new

template and coronally resliced at 1x1x1 mm3 resolution. The left hippocampus was

traced by a research assistant from these coronal images using ANALYZE (version

10, Mayo Clinic) using a previously published protocol 2. This assistant was blind to

information of the subjects’ temporal and grouping status for all measurements.

Hippocampal Longitudinal FSL-FIRST analysis: raw T1-weighted MRI scans were

firstly checked for obvious anatomical or positional abnormalities and extract the

brain. Next, bilateral hippocampi were segmented using Functional MRI of the Brian

(FMRIB)'s Integrated Registration and Segmentation Tool (FIRST, v5.0.0) in

FMRIB's Software Library (FSL, version v5.0). We used the vertex output of

hippocampi to assess the longitudinal morphometric change on the hippocampal

surface.

FreeSurfer (v5.1.0) to analyze the cortical structures: The longitudinal pipeline in

FreeSurfer was applied 3 to extract reliable longitudinal measures of cortical

thickness. It creates an unbiased within-subject template space and image 4 using

robust, inverse consistent registration 5. The surface map indicating the rate of

thickness changes were calculated between two time points for each subject. These

maps were used to test the longitudinal training effect on a vertex basis.

1

COGPACK Computerised Cognitive TrainingManual of Procedures

RationaleThe development of new intervention strategies in the prevention or delay of Alzheimer’s disease onset are an international priority given the health burden associated with the disease. Mental activity, or cognitive training, has recently been identified as a means to provide stimulating brain exercises to optimize the potential for neural plasticity: maintaining and potentially improving brain function.

The suite of cognitive training exercises are computer-based in keeping with recent commercial and research programs (e.g. Mahncke et al 2006; Schreiber, 1999; Hofman et al 2003) and according to the literature, multimodal, multi-domain and task load-graded training in the areas of memory, executive function, attention and speed of information processing.

CogpackThe Cogpack package of exercises were developed by a German neuro-rehabilitation centre and are largely based upon well established neuropsychological tests and principles and, moreover, their effective use with psychiatric patients has been reported (McGurk et al; 2005).

Fourteen exercises will be administered in the SMART study with multiple exercises being used to target the one generic cognitive domain and encompassing both verbal and visuo-spatial stimuli in order to maximize cognitive and neurological outcomes. Within-domain multiplicity is also required to ensure participant engagement and avoidance of ceiling effects during the long 6-month training phase.

Table 1: Cogpack exercise, cognitive domain, description.

Cogpackexercise name Cognitive Domain

Neuropsychological Description

Reading Verbal episodic memory Recognition multiple choice recall format of structured verbal information

Memory: Names

Verbal episodic Recognition recall of verbal information

Memory: Shopping List

Verbal episodic Recall

Sequence: Months

Verbal working memory Tracking and holding verbal information

Memory: Traffic

Visual episodic Recognition recall format visual information

Memory: Forms

Visual episodic Recognition recall format visual information

Route Visuospatial working memory

Free recall visuo-spatial information

Reaction: 3x2 Speed of processing Speed of comprehension.UFOs Speed of processing Reaction time task to visual stimuli

2

Anagrams Executive Deductive word completionLogic: Blocks Executive Reasoning pattern completion.Logic: And UK Executive Deductive reasoning and

comprehension taskSearch Attention Visual scan & searchConnect Attention Complex tracking and shift

Twelve of the 14 exercises require a simple touch-screen response with no mouse operation and in this way will avoid training difficulties in the computer-naïve; furthermore the touch-screen medium has been shown to be efficacious in cognitively impaired elderly (Hofman et al 1996). Two exercises have been chosen to specifically train mouse-based speed-dependent responses, and these will be completed before commencement of the two additional exercises that require mouse skills. The exercise set has been predetermined for each separate training session and will be randomly delivered to each participant.

ProcedureParticipants will begin their first session with an introductory session including the provision of information regarding the training and obtaining their user code. A maximum of ten participants will complete training within the same session. Each session will involve the research assistant setting up the participant at their computer station, entering the participant code and opening the specific session of exercises. The RA will then provide computer support and positive feedback to maintain motivation and compliance. At ten minute intervals the RA will ensure the participant has started the next training exercise to guarantee that an exercise is completed for each domain at every session.

Cognitive training feedbackDuring any small group or individual training, discussion and motivational support and advice will naturally occur. In order to standardize this feedback throughout the SMART trial a series of ‘strategies’ (internal cognitive strategies based upon well established rehabilitation techniques) have been developed for each exercise. This feedback is to be provided to all participants.

Administration

First training session1. Participant File codesEach participant of the SMART cognitive training exercises needs to be entered into the Cogpack system- a person file code. There are three stages to identifying the person- 1- ID codes which can only contain numbers. Then name, and last is Your code which can be numbers, letters or symbols. Keep same as ID code for tracking.

1. Open Cogpack2. Open [person] from menu3. go to [input]4. complete all fields with information from participant file5. Enter participant ID code- must be numbers only6. Enter first name

3

7. click [both of the above] okay 8. Enter participant ‘Your code’ can be anything

2. Welcome Welcome each participant to the mental exercise training program.

“This is the beginning of your mental activity program. The program will last for 52 sessions over 26 weeks, three sessions each week. Every session will consist of 5 different computer exercises for memory, problem solving, speed, and attention, and will take about 50 minutes to complete”.

“Each exercise has been selected to be challenging so you are having trouble don’t worry, just continue to complete the task. Everyone will complete the same program but in a different sequence, so don’t be concerned if other people are completing different exercises to you”.

“I will be here during your exercise session to start you off and to help you, so if you are having any concerns just ask. Let us begin your first session”.

3. Start trainingCogpack commands:

1. open Cogpack2. click on [person]3. go to [code], and enter participant’s ID code4. click on [training]5. click on [series]

4

6. click on session # for that participant7. [start] the exercise session will proceed automatically

Using Equipment

Participants can use the mouse, touch screen or keyboard. Please identify with each participant which method/s suit them best. Get participant to practice on touch screen. They must hit screen in single movement with finger pad NOT nail, and hit on or above icons.

Progress through the exercisesEvery ten minutes insure that the participant has moved on to the next exercise so that they can complete 5 exercises within each session.

At the end of each exercise participants can review their progress by clicking on [compare your scores]. Only participant codes will be evident with their performance being the last entry.

Exercise feedback MemoryReading (session 1; 4; 7; 10; 14; 16; 19; 22; 25) A passage is provided to read followed by multiple choice questions.Participants will complete 2 trials of each passage per session. After completing the first reading trail, click the [complete trial again] option. See below

Strategy: While reading the passage repeat the important information in your head- this is a rehearsal strategy to increase your learning.

5

Memory: Names (session: 2; 5; 8; 11; 13; 17; 20; 23; 26)A set of 4 names are to be learnt and then recognized from a set of 12. 10 trials.Strategy: Repeat the names in your mind as you read them, a rehearsal strategy. You could shorten the names to the first 3 letters and make a new ‘word’ – chunking the information together.

Sequence (session: 3; 6; 9; 12; 14; 18; 21; 24)Months of the year are displayed and the user has to determine if in order with the preceding item.Strategy: Repeat the items in your mind in order to ‘hold’ them between each display.

Memory: traffic (session: 1; 4; 7; 12; 15; 19; 22; 26)A set of 6 road traffic signs are displayed to be learnt, and then a set of 12. 10 trials.Strategy: Link the signs together between the images eg. Stop sign and train sign could become stop for a train- make associations between the information to assist your memory. Another strategy is to verbally describe the visual image in order to give you a visual and verbal memory of the information.

Memory: forms (session: 2; 5; 9; 10; 13; 16; 18; 20; 24)A set of 3 geometric shapes are displayed followed by a set of 12 items for recognition.Strategy: Link the signs together between the images eg. Stop sign and train sign could become stop for a train- make associations between the information to assist your memory. Another strategy is to verbally describe the visual image in order to give you a visual and verbal memory of the information.

Memory: route (session: 3; 6; 8; 11; 15; 21; 23; 25)A journey through a map is displayed and user has to recall trail.Strategy: Describe the trial in verbal commands to yourself- go left 3 down 2 left 2, and repeat.

Memory: shopping (session: 3; 6; 8; 11; 15; 21; 23; 25)List of shopping items to recall.Strategy: Chunking items, and repeat.

Speed of processingReaction: 3x2 mouse (session: 4; 7; 11; 14; 18; 25)Follow command.

Visuomotor: raft (session: 5; 8; 10; 13; 16; 19; 24)Move target through screen.

UFO’s: permanently adaptive (session: 1; 3; 6; 17; 20; 23; 26)Hit targets as soon as they appear on screen.Falling Stars (session: 2; 9; 13; 19; 22; 26)Catch items.

6

Strategy: Keep finger or mouse in centre of screen and move it in a calm systematic manner to increase accuracy.

ExecutiveAnagrams (session: 1; 9; 13; 19; 22; 26)Re-ordering letters to spell known words.Strategy: Start with a single consonant and generate options, then progress to another letter

Guess-words (session: 2; 5; 12; 16; 21)Following clues to deduce the target word.Strategy: Use your general knowledge, systematically try letter options- eg start with vowels.

Labyrinths (session: 3; 8; 11; 18; 23)Maze completionStrategy: Take your time and plan your route

Logic: blocks (session: 4; 7; 14; 17; 24)Complete patternsStrategy: Recognise what the pattern is by systematically thinking about how / why the items are in that pattern.

Logic: And UK (session: 6; 10; 15; 20; 25)Comprehending a statement and applying the rule to an array of pictures.Strategy: Read the statement carefully so you understand what it means. Apply the rule and check your work.

AttentionSearch (session: 1; 3; 5; 7; 9; 11; 13; 15; 17; 19; 210; 22; 24)Visually scanning to locate a target itemStrategy: Look along each row or section in a systematic left right motion.

Connect: I A @ B (session: 2; 4; 6; 8; 10; 12; 14; 16; 18; 21; 23; 25; 26)Linking numbers and lettersStrategy: Recite items to help keep track of your progress

End Session

Thank participant for coming to the session and remind them when you will see them next.

7

Functional MRI results

Posterior Cingulate (PC) seed/ Default Mode Network (DMN)

The sole RES x time effect was to decrease FC between PC and left inferior temporal

lobe (ITL_L, df=67, F=14.8, p<0.001, Figure S1A) as well as ACC†1 (df=67, F=23.3,

p<0.001, Figure S1A).

The COG x time interaction also decreased the FCs of PC and right ACC (ACC_R†,

df=67, F=13.9, p<0.001, Figure S1B) and left superior frontal lobe (SFL_L†, df=67,

F=31.7, p<0.001, Figure S1B).

The RES x COG x time interaction indicated that the combination of the two types of

training significantly reduced the FC between PC and two regions, ventral ACC†

(df=65, F=5.3, p=0.017, Figure S1C) and within the PC† itself (df=65, F=6.1,

p<0.001, Figure S1C).

In general, training has a global effect on decreasing the FCs of the PC (Figure S1).

1† Where marked (†) in text or figures indicates statistical results after additional control for baseline FC, because of significant differences on baseline FC between groups.

8

Figure S1. Longitudinal training effects on the Posterior Cingulate (PC) seed FCs. PC

seed is shown at the right top corner. Regions with whole brain cluster level pFDR-

corrected<0.05 were considered significant. RES x time had a longitudinal effect on

decreasing FC network to the left ITL as well as ACC. COG by time also decreased

the PC FCs to right ACC (ACC_R) and left superior frontal lobe (SFL_L). RES x

COG x time indicated that the combination of the two types training significantly

reduce the FC between PC and two regions, ACC and PC itself. Several results further

remained significant after controlling for baseline connectivity (due to significant

baseline group difference on FC), indicated by † at p-value

9

Hippocampal (HP) Seed

For the RES x Time interaction, functional connectivity between HP and right inferior

temporal lobe (ITL_R†2 df=67, F=18, p<0.001, Figure S2A) was significantly

decreased. By contrast, the PRT x time interaction was also significant, but in the

opposite direction when examining FC between HP and right middle frontal cortex

(MFC_R, df=67, F=13.0, p=0.001, Figure S2A).

The CT x time interaction produced positive changes on the hippocampal FCs of right

inferior temporal lobe (ITL_R†, df=67, F=10.6, p=0.006, Figure S2B) as well as left

superior frontal gyrus (SFG_L, df=67, F=6.1, p=0.012, Figure S2B).

The three-way interaction found significant areas at the right superior frontal gyrus

(SFG_R†, df=65, F=7.0, p<0.001) and dorsal ACC (df=66, F=4.6, p=0.005). Post-hoc

analysis showed that the combination of CT and PRT increased the FCs between HP

and these two regions more than the stand alone interaction (Figure S2C).

2 Where marked (†) in text or figures indicates statistical results after additional control for baseline FC, because of significant differences on baseline FC between groups.

10

Figure S2. Longitudinal training effect on hippocampal (HP) FCs. Bilateral

hippocampal seed is shown at top right corner. Regions with whole brain cluster level

pFDR-corrected<0.05 were considered significant. RES x time decreased FC between HP

and right inferior temporal lobe (ITL) but increased the FC between HP and right

middle frontal cortex (MFC). COG x time elevated hippocampal FCs to the right ITL

and left superior frontal gyrus (SFG_L). RES x COG x time indicated that the

combination of the two types of training significantly increased the FC between HP

and two regions, right SFG and ACC. Several results further remained significant

after controlling for baseline connectivity (due to significant differences between

groups on baseline FC), indicated by † at p-value.

11

1. Suo C, Leon I, Brodaty H, Trollor J, Wen W, Sachdev P et al. Supervisory experience at work is linked to low rate of hippocampal atrophy in late life. Neuroimage 2012; 63(3): 1542-1551.

2. Valenzuela MJ, Sachdev P, Wen W, Chen X, Brodaty H. Lifespan mental activity predicts diminished rate of hippocampal atrophy. PLoS One 2008; 3(7): e2598.

3. Reuter M, Schmansky NJ, Rosas HD, Fischl B. Within-subject template estimation for unbiased longitudinal image analysis. Neuroimage 2012; 61(4): 1402-1418.

4. Reuter M, Fischl B. Avoiding asymmetry-induced bias in longitudinal image processing. Neuroimage 2011; 57(1): 19-21.

5. Reuter M, Rosas HD, Fischl B. Highly accurate inverse consistent registration: a robust approach. Neuroimage 2010; 53(4): 1181-1196.

12