Figure 1.
Face recognition memory for faces and shoes.
(A) Compared to controls (O), CPs (X) showed worse recognition performance after one year for target faces but not for shoes. (B) The difference in face recognition is largely due to a decreased false positive rate among CPs who failed to detect the target faces more often than controls.
Figure 2.
(A) Compared to controls, CPs showed worse recognition performance in free name recall. (B) Performance is influenced by confounding factors, e.g. age. (C) After accounting for confounding factors, performance differences between CPs and controls are more pronounced.
Figure 3.
Comparison of famous face test (BFFT) and one-year recognition memory.
Overall differences between CPs (X) and controls (O) are more pronounced in the test of one-year recognition memory than in the BFFT. Among controls the test results seem dissociated, with participants scoring bad in either of the two tests but not in both. Several controls who scored low in the BFFT, but all of them show normal performance in the standardized one-year recognition memory test.
Figure 4.
Overview of standardized test performance.
Clustering participants according to their standardized test performance distinguishes CPs (top rows, marked by initials) from controls (bottom rows, marked by small circles) and reveals a dissociation of the deficits in CP into three homogeneous subtypes. In the table individual z-scores (color code at the bottom) are displayed for all of the tests (columns) based on estimated residuals for each participant. Significant deficits (modified t-test, 5% cutoff) are marked by an asterisk, missing values are denoted by X. Columns on the left side (Summary Z-scores) display aggregate performance across all or several tests, columns on on the right side display performance in individual tests (Separate Z-scores). Rows are ordered according to an unsupervised clustering on aggregate test results (complete linkage analysis). Abbreviations: RT = reaction time, PT80% = presentation time needed for 80% correct performance, BFFT = Bielefeld famous face test.
Figure 5.
Correlation of experimentally tested deficits and self-assessment.
Scores in questionnaire based self-assessment and overall residuals in experimental tests of face recognition performance are correlated in CPs (ρ = −0.55, p = 0.0345).