Fig 1.
(a) depicts the ambiguous Necker lattice, a variant of the Necker cube [5]. The unambiguous variants are displayed in (b) with the front side facing towards the top (front-top = FT) on the left and the front side facing towards the bottom (front-bottom = FB). Stimuli were created in the laboratory of Dr. Kornmeier.
Fig 2.
Stimuli were presented in pairs one after the other and formed one Observation Sequence (OS). Each stimulus was presented for 800 ms. Stimulus 1 (S1) and Stimulus 2 (S2) were presented in succession and temporally separated by an inter-stimulus interval with a dark screen for 400 ms. Presentation of a dark screen for 1000 ms separated subsequent OS from each other. The experimental paradigm consisted of two tasks: during the presentation of the lattice stimulus S1 participants indicated the perceived orientation of S1 (Task 1). During the subsequent presentation of the lattice S2, they compared their perceived S2 orientation with the previously perceived and memorized orientation of lattice S1 and indicated either percepts of identical or reversed orientation (Task 2). Notice that the Task 1 was only related to stimulus S1. Neither the information about the preceding and subsequent stimuli nor information about their ambiguity levels was necessary for the execution of this task.
Fig 3.
The current experiment consisted of four separate experimental conditions with a 2 x 2 design. SU(CU): Both lattices S1 and temporal context stimuli (preceding and upcoming S2) were unambiguous; SU(CA): S1 unambiguous and temporal context ambiguous; SA(CU): S1 ambiguous and temporal context unambiguous; SA(CA): both S1 and temporal context ambiguous. Ambiguity levels of S1 and the temporal context were kept constant and were thus highly predictable within conditions but differed between conditions. In Experiment 1, each experimental block consisted of 180 observation sequences (OS) with the same stimulus pairs. Each block was repeated 3 times across the experiment. (U = Unambiguous, A = Ambiguous).
Table 1.
Number of trials of Experiment 1.
Fig 4.
Reversal rate Necker lattices.
We calculated the reversal rate towards front-bottom (FB) and the reversal rate towards front-top (FT) views of the Necker lattice from the preceding S2 of the previous pair towards the currently seen S1. A schematic overview of condition SA(CA) can be seen in a) and an example with stimuli in b). In c) the average (±SD) values are displayed for both view orientations and for each condition.
Fig 5.
Reaction time data for task 1.
Blue colours indicate reaction times to ambiguous stimuli SA (a) and red colours to unambiguous stimuli SU (b). Reaction times show opposite effects of stimulus ambiguity within the temporal context for ambiguous compared to unambiguous currently observed stimuli: reaction times were generally shorter when the stimuli S2 from the temporal context were of the same ambiguity level as the currently perceived stimulus compared to those conditions with differing ambiguity levels of temporal context stimuli S2 compared to the perceived S1. (c) List of median reaction time [s] values with the interquartile ranges [s], separately for each experimental condition.
Fig 6.
ERP effects of sensory quality within the temporal context.
(a) ERP traces at electrode Cz during perception of an ambiguous lattice SA, when the stimuli in the temporal context (i.e. S2 from the preceding pair and the predicted S2 from the current pair) were unambiguous (“SA(CU)”, dark blue continuous trace) and when the stimuli from the temporal context were ambiguous (“SA(CA)”, light blue dashed trace). Notice that the same ambiguous current SA lattice stimulus evoked larger P200 and P400 amplitudes with an unambiguous temporal context compared to an ambiguous temporal context. (b) Voltage maps (top) showing the spatial distribution of P200 (left, t = 214 ms) and P400 (right, t = 326 ms) and scatter plots (bottom) showing individual mean amplitude data for the P200 (left) and the P400 (right), which correspond to (a). Notice that for almost all participants the P200 and P400 ERP components evoked by SA show larger amplitudes when the temporal context stimuli were unambiguous (data points are above the diagonal). (c) ERP traces during perception of an unambiguous lattice SU, when the temporal context stimuli were unambiguous (“SU(CU)”, dark red continuous trace) and when the temporal context stimuli were ambiguous (“SU(CA)”, light red dashed trace). Notice that the amplitude of the P400 evoked by the one and the same unambiguous present SU lattice stimulus varied as a function of the ambiguity level within the temporal context. No such effect is visible for the P200. (d) Same logic as in (c) but with an unambiguous present stimulus SU (data related to c; Voltage maps: P200—t = 222 ms, P400—t = 358 ms). U = unambiguous, A = Ambiguous, S = stimulus S1, C = temporal context (preceding and subsequent S2).
Fig 7.
The symbolic announcement in Experiment 2.
Ambiguous stimuli are announced by question marks, while unambiguous stimuli are announced by exclamation marks. Before each block of condition SA(CU) a picture was presented with a question mark symbol (left) pointed towards an exclamation mark symbol (right). Before each block of condition SA(CA) a picture with two question marks was presented.
Fig 8.
Experiment 2 consisted of two separate experimental conditions (bottom part). Within experimental conditions, ambiguity levels of currently observed S1 stimuli and ambiguity levels of their temporal context (preceding and subsequent S2) were kept constant and were highly predictable within conditions. The ambiguity level of the temporal context differed between conditions. SA(CU): current ambiguous stimulus SA and unambiguous temporal context; SA(CA): both SA and temporal context ambiguous. One observation sequence (OS) consisted of an S1 stimulus (800 ms), an inter-stimulus interval (400 ms), a S2 stimulus (800 ms), and an inter-observation sequence interval (1000 ms). One experimental block consisted of a symbolic announcement and three repetitions of the observation sequence. Each block of Exp. 2 was followed by an experimental block from another unrelated experiment with completely different stimuli to increase the temporal distance between blocks of Exp. 2 and thus to minimize low-level memory effects (e.g. priming, adaptation, etc.) from one block to the other. (U = Unambiguous, A = Ambiguous).
Table 2.
Number of trials of Experiment 2.
Fig 9.
Reaction time results from Experiment 2.
Bottom row depicts the median reaction times evoked by a currently observed stimulus separately for the two experimental conditions with unambiguous (black) and ambiguous (grey) stimuli in the temporal context of SA. Reaction times are separately shown for the first (circles, first column), second (squares, second column), and third pair (diamond, third column). For comparison, the results from Exp. 1 (triangles) are plotted on the right (fourth column). The data from individual participants are represented with small icons, while the large icons represent the median reaction time data with interquartile ranges (whiskers). In the middle row the median reaction time differences (large icons ± interquartile ranges) between the two conditions (unambiguous temporal context minus ambiguous temporal context) are depicted together with data from the individual participants (small icons). The top row shows the sizes of the reaction time effects (res). U = Unambiguous, A = Ambiguous.
Fig 10.
Grand mean ERP results from Experiment 2.
Grand means (±SEM) in response to a currently observed stimulus are separately shown for conditions SA(CU) (dark blue solid lines) and SA(CA) (light blue dotted lines). SA-evoked ERP traces from stimulus pair 1 (left), 2 (middle), and 3 (right) are depicted separately. All traces are displayed for electrode Cz.
Fig 11.
P200 ERP results from Experiment 2.
Bottom row depicts the mean amplitudes of the P200 evoked by a currently observed stimulus SA separately for conditions with unambiguous (CU, black) and ambiguous (CA, grey) stimuli in the temporal context of SA. P200 amplitudes are separately shown for SA from the first (circles, first column), second (squares, second column), and third pair (diamond, third column). For comparison, the results from Exp. 1 (triangles) are plotted on the right (fourth column). The data from individual participants are represented with small icons, while the large icons represent the mean amplitudes with SEM (whiskers). In the middle row, the mean ERP differences (large icons ±SEM) between the two conditions (unambiguous temporal context minus ambiguous temporal context) are depicted together with data from the individual participants (small icons). The top row shows the effect size (res) of the temporal context effects. U = Unambiguous, A = Ambiguous.
Fig 12.
P400 ERP results from Experiment 2.
Bottom row depicts the mean amplitudes of the P400 evoked by a currently observed ambiguous stimulus SA, separately for conditions with unambiguous (CU, black) and ambiguous (CA, grey) stimuli in the temporal context. P400 amplitudes are separately shown for the first (circles, first column), second (squares, second column), and third pair (diamond, third column). For comparison, the results from Exp. 1 (triangles) are plotted on the right (fourth column). The data from individual participants are represented with small icons, while the large icons represent the mean amplitudes with SEM (whiskers). In the middle row the mean ERP differences (large icons ±SEM) between the two conditions (unambiguous temporal context minus ambiguous temporal context) are depicted together with data from the individual participants (small icons). The top row shows the effect size (res) of the temporal context effects. U = Unambiguous, A = Ambiguous.