Scan Patterns Reveal Strategic Influence on Raven (Hayes & Petrov, 2011)

Successor Representation Scanpath Analysis (SRSA) was used to extract individual differences in eye movement patterns and predict cumulative score across two separate test sessions on Raven’s Advanced Progressive Matrices.  The SRSA method extracted the underlying structure from complex patterns of sequential eye movements revealing two readily interpretable components of performance.

SRSA_sum1
Systematic row-wise scanning (left) was associated with higher scores, while toggling (right) between the problem matrix and response area was associated with lower scores.

A “systematic” component which is characterized by a prominent diagonal box structure.  The 3 x 3 red boxes indicate the benefits of systematically scanning within a given row of the problem matrix.  The positive red values “dripping” from each box indicate the systematic integration as participants moved from row to row.

The second was an “anti-toggle” component indicated by the strong blue activity across the response area.  The anti-toggle component indicates that participants who made fewer toggles from each cell of the problem matrix to the response area achieved higher scores than participants who toggled more frequently.

SRSA_sum2
Together the systematic and anti-toggle components predicted over 40% of the variance in Raven score.

These two components of strategy predicted 56% of the variance in cumulative Raven scores.  Leave-one-out cross-validation produced the same two components (systematic and anti-toggle) and still predicted 41% of the variance in RAven scores.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s