Bachelorproject
Is Consciousness a Gradual or Dichotomous Phenomenon:
A Comparison Between Techniques and Subjective Measurements
Student: Saskia Vinke Student number: 10782869 Supervisor: Timo Stein Date: June, 2017
Abstract
Whether there is a strict dissociation between conscious and unconscious processing is a much-debated issue. Evidence for consciousness as a gradual process or an all-or-none process has been found in several studies using different kind of techniques and measures. In this study techniques and measures were compared under identical circumstances so that objective conclusions could be drawn. The Attentional Blink and Backward Masking were both measured with the Continuous Scale and the PAS. The results showed a an All-or-None shift for the Backward Masking measured with the Continuous Scale, for the other experiments the distribution was gradual. Therefor it can be concluded that it is most likely that the shift from unconscious to conscious is a gradual one.
Introduction
The problem of consciousness, otherwise known as the mind-brain problem, was originally the realm of philosophers. The discussion was mainly focused on whether you were a dualist or a materialist and a common definition of the meaning of consciousness was still lacking (Gazzaniga, Ivry, & Mangun 2014). In the 1995 edition of the International Dictionary of Psychology, the psychologist Stuart Sutherland defined consciousness as the having of perception, thoughts, and feelings; awareness. To be conscious it is only necessary to be aware of the external world. A subpart of the wide concept consciousness is visual consciousness, and this is becoming a large subject in consciousness research. It is of fundamental importance to understand the degree to which stimuli are processed without conscious awareness, without this understanding we cannot understand our own visual system.
In the past twenty years, the fields of cognitive science, neurophysiology and brain imaging have mounted a solid empirical attack on consciousness. As a result, the problem has lost its speculative status and become an issue of experimental ingenuity (Dehaene, 2014). When we lay our eyes on an object we immediately see it –and become aware of its shape, colour and identity, this made consciousness difficult to experiment within cognitive science. In the past twenty years, cognitive scientists have discovered an amazing variety of ways to manipulate consciousness. By flashing words so briefly that viewers will fail to notice it, distracting the attention of the viewer or letting the brain do the magic by presenting two distinct images to each eye we can manipulate our consciousness and test it (Dehaene, 2014). Whether there can be a strict dissociation between conscious and unconscious processing is a debated issue. Both evidence for consciousness as a gradual process or an all-or-none process has been found in several studies.
Baars (1997) introduced the global workspace theory. This theory is easiest to explain in terms of the “theatre metaphor”. In this so-called “theatre of consciousness” there is a spotlight that shines on the stage, this spotlight is your selective attention. The bright spot on the stage reveals the actors interacting with each other; this stands for our contents of consciousness. The audience is not lit up-it is in the dark watching our play and therefore representing our unconsciousness. The global workspace theory is in line with the ideas that
consciousness is an dichotomous phenomenon because the shift of the spotlight is not gradual, the light is either on it or not. This theory is opposite of the idea that the shift from unconscious to conscious is gradual.
There are different measures for consciousness; this paper will mainly focus on the subjective measures. These measures require subjects to report their own mental states. Most simply, measures have been
used to ascertain whether a person knows that they know (Seth, Dienes, Cleeremans, Overgaard, and Pessoa, 2008).
Conscious awareness of different cognitive processes has often been investigated with so called subjective measures of consciousness (Dienes & Seth, 2010; Pessoa, Japee, Sturman, & Ungerleider, 2006; Sanberg, Timmermans, Overgaard, & Cleeremans, 2010; Szczepanowki, 2011; Szczepanowski & Passoa, 2007; cited in Szczepanowski, Taczyk, Wierzchon, & Cleeremans, 2013).
For example in the study of Sergent and Dehaene (2004) participants were asked to evaluate the visibility of target words on a continuous scale. The study showed that the participants used this continuous scale in an all-or-none fashion. The continuous scale is a purely introspective measure by which participants are asked to indicate the level of object visibility. The rating varies on a scale between “no experience” and “absolute clear image”. The participants choose any point on the continuum between the ends of the scale (Wierzchoń,
Asanowicz, Paulewicz, & Cleeremans, 2012). There is a significant correlation between the perceptual identification task performance and the awareness (Sandberg et al., 2010, and Sergent & Dehaene, 2004). Sergent and Dehaene (2004) revealed a all-or-none pattern from their results, this was interpreted as an argument for the dichotomous character of conscious perception.
In 2004 Ramsøy and Overgaard asked their participants to describe the quality of their visual experience and to use a scale they could create themselves. The experimenters suggested that the scale started with ‘no experience’ and ended with ‘a clear image’. All five participants ended up using a 4-point scale with the elements (1) ‘No experience, (2) ‘Brief glimpse’, (3) ‘Almost clear image’, and (4) ‘Absolutely clear image’. In the study of Overgaard, Rote, Mouridsen, and Ramsøy (2010) the PAS showed that conscious perception is gradual. One advantage of PAS is that the scale is directly asking what the experimenter is looking for, but at the same time this can be seen as a disadvantage because it depends on how good participants are at reporting their experience (Sanberg, Timmermans, Overgaard, & Cleeremans, 2010).
Condifence ratings can be used in two different manners; (1) participants directly report their confidence in having perceived something, and (2) by reporting their confidence in having provided a correct answer (Bernstein & Eriksen, 1965 and Cheesman and Merikle, 1984, cited in Sanberg, Timmermans, Overgaard, & Cleeremans, 2010). In contrast with the PAS there are different scales for confidence ratings, although most variations include ‘guessing’ or ‘no confidence’ in the description of the lowest rating. A disadvantage of PAS was that people might no be so good at introspection; in this case confidence ratings may
have an advantage, as participants are not asked directly to introspect. An disadvantage in participants rating their own performance may be that different participants will use different criteria to decide their confidence (Sanberg, Timmermans, Overgaard, & Cleeremans, 2010).
One of the most recent measures of conscious content is Post-decision wagering. After performing the task, the participants can place a wager on having performed the task correctly. The wagers are based on the awareness of the participants, but they never have to report their awareness in an introspective way. For this reason Persaud et al. (2007) put decision wagering forward as a direct measure. There is also sad that post-decision wagering is an intuitive method (Sanberg, Timmermans, Overgaard, & Cleeremans, 2010). Schurger and Sher (2008) reported that only two out of more than hundred participants that were tested with PDW adopted an optimal strategy, even when encourage to wager high.
So for PAS the participants report awareness directly, for CR they report their confidence in being correct, and for PDW participant wager on being correct.
The experiments from Sandberg et al. 2010 showed that the use of different subjective measures could lead to different results. In the experiment scales were compared between groups, these subtle differences between the groups could be seen as an cause for the differences. However, due to the size of the groups and the fact that the groups did not show difference in age nor task accuracy, any impact of group differences seems modest.
In the experiment from Sandberg et al. (2010) the data suggested that PAS appears to be more sensitive to different levels of awareness, each corresponding to different accuracy levels. Confidence ratings fell somewhere in between PAS and PDW.
Szczepanowski, Taczyk, Wierzchon, and Cleeremans (2013) also compared the three different measures of awareness. They found that the confidence measure was the most graded. The results from the perceptual awareness scale and post-decision wagering suggested that consciousness was perceived in a dichotomous matter. The results of these studies can suggest that all of these different measures are not all measuring people’s perceptual awareness in the same way.
The most recent used techniques that are being used with the previous discussed scales are backward masking, attentional blink, and breaking continuous flash suppression.
Visual backward masking occurs whenever the visibility of one stimulus, called the target, is reduces by a mask presented right after the target (Breitmeyer & Ogmen, 2000). After the masking the participants report the visibility of the target. In the study of Del Cul, Baillet, and Dehaene (2007) subjective visibility ratings neatly clustered into two well-defined “seen” and “not-seen” states, this is evidence for an all-or-none phenomenon.
In the 2006 study of Overgaard et al. conscious visual perception was also studied with a masking task. But this time the results showed that there are different levels of consciousness perception and you cannot define them in terms of all or none.
Sergent and Dehaene (2004) studied the use of the continuous scale in a classical attentional blink paradigm in which the stimulus-onset asynchrony between target 1 and target 2 was varied. The attentional blink degrades the available information on target 2. If consciousness is gradual the distribution of participants’ responses would be expected to shift gradually from the low end of the scale when the attentional blink when the attentional blink was strongest. However, when the result shows two types of peaks, not-seen and seen, it would appear that consciousness is all-or-none. Sergent and Dehaenes study showed that the last option was the case and the target was seen in an all-or-none fashion.
In a second experiment, In the same Sergent and Dehaene also tested the backward masking task with the continuous scale, the responding on this task was not all-or-none but gradual.
Nieuwenhuis and de Kleijnn (2011) also used the attentional blink experiment to predict whether there is a gradual change or a discontinuous transition between unconscious an conscious perception. The four attentional blink experiments from their study showed the subjects used the consciousness rating scales in a continuous fashion. These results show that the all-or-none is not always the case using the attentional blink.
To analyse all the differences between different tasks and scales, the results of all the tasks should be directly compared in one study. In this paper the effect of the PAS and continuous scale will be compared by using both the Attentional Blink paradigm as well as the backward Masking Paradigm. By using the different scales and techniques in the same controlled conditions with the same stimuli conclusions on the working of these different scales and techniques can be drawn. If both the scales measure the subjective transition from unconscious to conscious perception the outcome, either gradual or all-or-none, should not differ between scales. Therefor the first hypothesis is that all the scales should have the same outcome, either gradual or all-or-none.
When the first hypothesis does not occur, you can state that one or both of the scales are not measuring subjective visual consciousness. To make sure this is the reason for the differences between the scales a second hypothesis will be looked at. This second hypothesis states that within the two different techniques the outcomes of the scales should not differ. This means that for the Attentional Blink the outcome of the PAS and Continuous Scale are telling the same story; this is the same for the Backward Masking Paradigm.
Methods
Experiment 1: An attentional blink paradigm and backward masking measured with the continuous
scale
Subjects
A total of 16 subjects, of which 1 were left-handed and 16 right-handed, participated in experiment 1. Of the 16 participants, 7 were women and 9 men; age ranging from 19 to 24 with a mean of 21.1 and a standard deviation of 1.6. All of the participants have been acquired through www.lab.uva.nl of which the majority were first-year psychology students, which got rewarded with participation points. The rest of the participants have been recruited differently and have not received any reward. All had normal or corrected-to-normal vision.
Design
Each subject participated in both the attentional blink task as well as the backward masking task. In both methods the subjective measurement was the continuous scale.
Materials
Both tasks were presented on a black background at the centre of the computer screen (100-Hz refresh) using Toolbox for MatLab software.
The attentional blink paradigm consisted of 384 trials. In the experiment three scrambled faces were showed next to each other (see Fig. 1). The scrambled faces consisted of 24 different stimuli that were randomly presented for 100 ms. T1 consist of the same three faces except that the middle face is filled with either green diamond or circles (see Fig. 2). The position of T1 was somewhere in between the 8th and 12th stimuli. After T1
is showed the three faces become scrambled again. T2 shows the three faces again, but now it is the case that there is an complete unscrambled face either on the right or left (Figure 3). There are three different lags for T2 that are shown after T1; lag 1, lag 3, and lag 8.
Figure 1. Three scrambles faces next to each other in the Attentional Blink
Figure 2.T1: On the left side three scrambled faces with a green diamonds and on the right side scrambled faces with circles in the middle.
Figure 3. T2: Two scrambled faces and a whole face either on the left or right side.
The backward masking paradigm consisted of 384 trails. Each trial begins with a fixation cross and after that a face on the left or right side of the cross (see figure 4). There are three different duration times for the faces; 10 ms, 20, and 20 ms with 10 ms blank. After the face was displayed a mask will be represented for 20 ms (see Fig. 4).
The subjective measurement was a continuous scale, consisting of 240 points, in which the subject could move over the bar with the mouse.
Figure 4. On the left the face shown in the Backward Masking Task and on the right the scrambled background
Procedure
The attentional blink and backward masking was counterbalanced between the subjects, so it differs between subjects which task they started with.
In the attentional blink paradigm participants initiated the task by pressing any key on the keyboard. Trials began with a fixation cross on the screen for 1 second, after this there was a blank for 0.5 second, followed by T1 and T2. After T2 subjects were asked whether T1 showed diamonds or circles, after their response the participant received feedback in the form of a green gross (correct) or a red cross (incorrect). After the objective measure the subjects must guess whether the face was located on the left or right side of the screen by pushing either the left or right arrow on the keyboard. At the end the continuous scale appeared in which participants could rate the face visibility of T2 on the continuous scale.
In the backward masking task the task was also initiated by pressing any key on the keyboard. Trials began with a fixation cross on the screen for 1 second, after this there was a blank for 0.5 second, followed by a face. After this the face location test is shown (left or right arrow key) and the face visibility test by using the continuous scale.
Experiment 2: An attentional blink paradigm and backward masking measured with the PAS
Subjects
A total of 10 subjects, of which 2 were left handed and 8 right handed, participated in experiment 2. Of the 10 participants, 4 were women and 6 men; age ranging from 19 to 27 with a mean of 22.8 and a standard deviation of 2.7. All of the participants have been acquired through www.lab.uva.nl of which the majority are first-year psychology students who got rewarded with participation points. The rest of the participants have been recruited differently and have not received any reward. All had normal or corrected-to-normal vision.
Design
Each subject participated in both the attentional blink task as well as the backward masking task. In both methods the subjective measurement was the PAS.
Materials
For description of the attentional blink and backward masking paradigm see Experiment 1. The subjective measurement was the PAS, which consisted of 4 options. The options on the PAS were; ‘1 = no experience’, ‘2 = brief glimpse’, ‘3 = almost clear experience’, and ‘4 = clear experience’.
Procedure
The attentional blink and backward masking was counterbalanced between the subjects, so it differs between subjects which task they started with.
For a description of the procedure of the attentional blink and backward masking paradigm see Experiment 1. After each trials the participant will rate the face visibility by using the different keys corresponding to the different PAS scale points (1, 2, 3, 4).
Data-analysis
The data-analysis each combination of technique and scale will be conducted in the same way, each in its one section. For every combination the same analyses will be tested.
For the different presentation times/lags the means and standard deviations of the PAS and Continuous scores will be calculated. The independent variable being presentation time/lag and the dependent variable PAS/Continuous score. A Repeated- Measures ANOVA will be used to technique/scale for testing the influence on presentation time/lag on mean PAS/Continuous ratings.
The means and standard deviations of face localization accuracy for different presentation times/lags within the techniques/scales will be calculated. The independent variable being presentation time/lag and the dependent variable face localization accuracy. To test for each technique/scale the influence of presentation time/lag on face localization accuracy a Repeated-Measures ANOVA will be used.
presentation times/lags, 4 levels of PAS or 20 levels of Continuous Scale on proportion of visibility button presses.
Finally, the distributions of the PAS/Continuous scores between Attentional Blink and Backward Masking will be compared with a Repeated Measures-ANOVA with 2 techniques, 3 presentation times/lags, 4 levels of PAS scores or 20 levels of Continuous Scale scores on proportion of visibility button presses.
For each Repeated Measures ANOVA the assumption of sphericity will be tested using Mauchly’s test of Sphericity. If the assumption is violated a corrected degrees of freedom will be used.
Results
Attentional Blink measured with the Continuous Scale
All trials with an incorrect T1 response were discarded. To test for an effect of lag on the subjective measure, the continuous scale, the mean and standard deviation of the continuous scale scores of each lag has been calculated, see Table 1. Figure 1 shows that the continuous scale scores increase when the distance of the lag increases. The analysis shows that this increase in subjective score is significant.
A Repeated Measures ANOVA was executed for the effect of the different lags on the continuous scores. The dependent variable is the mean scores of the continuous scale and the independent variable the three
different lags. Mauchly’s test for sphericity indicated that the assumption of sphericity had been violated, χ2(2) = 6.177, p = .046, therefore Greenhouse-Geisser corrected tests are reported (ε = .74). The results show that the different lags have a significant effect on the continuous scale scores, F(1.47, 22.11) = 17.05, p < .001.
Table 1
Means Continuous Scale Scores and Standard deviations for the different lags.
Lag 2 Lag 3 Lag 8
Mean SD .309 .179 .332 .161 .389 .154
Figure 1. Means Continuous Scale Scores for lags 2, 3, and 8.
For the objective measures the means and standard deviations of the face localization accuracy for each different lag was calculated, see table 2. Figure 2 shows a clear increase of accuracy for each lag.
A Repeated Measures ANOVA was conducted testing the influence of presentation times on face localization accuracy. Mauchly’s Test indicated that the assumption has not been violated χ2(2) = 2.974, p = .226. The results show that the accuracy of the face localization improves with each different lag, F(2, 30) = 21.48, p < .001.
Mean T2 visibility rating for the different lags show a classic Attentional Blink pattern (see Figure 3).
Table 2
Means face localization accuracy and Standard deviations for the different lags.
Lag 2 Lag 3 Lag 8
Mean SD .691 .105 .751 .127 .817 .097
Figure 2. Means Face Localization accuracy for lags 2, 3, and 8.
For the distributions of the button presses of the Continuous Scale a Repeated Measures-ANOVA was conducted. The ANOVA shows that there is a significant effect of the interaction between lag and Continuous scores F(38, 570) = 5.606, p < .001. This means that the different lags have an influence on the distribution and the three visibility distributions differ from each other.
Figure 3 shows all the different distributions of the Continuous Scale scores for each lag (2, 3, and 8) in the attentional blink. All the lags clearly show that there are more presses at the first bin of the scale, this confirms the classical effect of the different lags in an Attentional Blink Paradigm.
The different graphs show that there is no clear all-or-none phenomenon for the continuous scale scores. Most of the button presses occurred in the first bin, for lag 2 this is 29%, for lag 3 23% and lag 8 18%. While there is no clear all-or-none distribution neither is the transition of the subjective visibility perfectly gradual, there is a peak at the first bin and in the last lag also at the last bin. From these different graphs it is hard to make any conclusions about whether subjective visual perception is dichotomous or gradual. What can be concluded is that in the Attentional Blink paradigm lag 3 is the most gradual lag and lag 8 the most all-or-none resembling.
Figure 3. Distribution of Continuous Scale scores for each lag in the Attentional Blink
Backward Masking measured with Continuous Scale
In table 3 are the mean scores and standard deviations of the continuous scale displayed for each presentation time. The mean scores show an effect for presentation time on continuous score. The longer the presentation time the higher subjects rate their subjective visibility, see Figure 4.
For the effect of presentation time on continuous scores a Repeated Measured-ANOVA was executed. Mauchly’s Test for Sphericity was used χ2(2) = 14.808, p = .001. Because the assumption is violated the Greenhouse-Geisser correction will be used (ε = .605). The Repeated Measure-ANOVA show a significant effect for presentation time on Continuous scores F(1.47, 22.11) = 17.05, p < .001.
Table 3
Mean scores Continuous Scale and Standard deviations for the different presentation times
10 ms 20 ms 20 ms + 10 ms blank Mean SD .206 .162 .323 .223 .551 .275
Figure 4. Means Continuous Scale Scores for the three presentation times
Table 4 shows the face localization accuracy and standard deviation for each different presentation time of T2. The means are illustrated in Figure 5 which illustrates a clear increase in correct responses on T2 when the presentation time of T2 increases.
To test the effect of presentation time on T2 accuracy a Repeated Measures-ANOVA was used. Mauchly’s Test of Sphericity indicated that the assumption of sphericity was violated χ2(2) = 8.018, p = .018, therefore degrees of freedom were corrected using Greenhouse-Geisser estimates of sphericity (ε = .696). There is a significant effect of presentation time on face localization accuracy F(1.393, 20.891) = 80.392, p < .001.
Table 4
Means face localization accuracy and Standard deviations for presentation time
10 ms 20 ms 20 ms + 10 ms blank Mean SD .679 .111 .770 .117 .917 .094
Figure 5. Means Face Localization accuracy for different presentation times.
To make conclusions about visual consciousness being a dichotomous or gradual phenomenon, we have to look at the different distributions for each presentation time. To test whether there is a significant difference between all of the distributions a Repeated Measure-ANOVA was used to look at the interaction effect of presentation time and Continuous scores. The ANOVA shows that there is an significant interaction effect for presentation time on score F(38, 570) = 14.045, p < .001.
Figure 6 shows that in the trials with the with shortest and longest presentation time most of the button presses occur in the first bin, so least visible. As the presentation times increases there is a shift toward the last bin, most visible. In the graph with the shortest presentation time 43,2 % of all presses were in the first bin as the presentation time extends, there is a shift in the distribution. The shift that occurs is the diminishing of the first peak and increase of the last bin, which is most obvious in the distribution of button presses in the 55 ms presentation time. In the graph with the longest presentation time 16,3% of all the presses were in the first bin and 24,3% in the last bin.
Figure 6. Distribution of Continuous Scale scores for presentation times in Backward Masking
To test the hypothesis whether the techniques are both valid measures for visual consciousness a Repeated Measure-ANOVA was used. An significant interaction effect between technique and presentation time/lag on Continuous score was found F(38, 570) = 6.946, p < .001.This effect indicated that there is a significant difference between the two techniques, Attentional Blink and Backwards Masking, in distribution. As shown in Figures 3 and 6 the distributions of the Attentional Blink Paradigm lean more towards a gradual transition, while the Backwards Masking scores are more distributed in an all-or-none fashion.
Attentional Blink measured with the PAS
Trials with an incorrect response to T1 were discarded. Table 5 shows the means and standard deviations of the PAS scores for each different lag. The means are illustrated in the plot in Figure 7 where you can see a clear increase in PAS score for each lag.
An analysis of variance (ANOVA) for repeated measures restricted to the T2 was conducted. Machly’s test for sphericity showed χ2(2) = 7.003, p = .030. Therefore degrees of freedom were corrected using
Greenhouse-Geisser estimates of sphericity. The Repeated Measures ANOVA revealed a significant main effect of lag on PAS scores F(1.25, 11.36) = 10.45, p = .005.
Table 5
Means PAS Scores and Standard deviations for the different lags.
Lag 2 Lag 3 Lag 8
Mean SD 1.504 .402 1.577 .452 1.733 .366
Figure 7. Means PAS Scores for lags 2, 3, and 8.
Table 6 shows the mean face localization accuracy and standard deviations for different lags within the Attentional Blink measured with the PAS. As shown in Figure 8 the different lags have effect on the face localization accuracy, as predicted for a Attentional Blink Paradigm.
A Repeated Measures ANOVA was used to look for the main effects of lag on face localization accuracy. Mauchly’s test for sphericity indicated that the assumption of sphericity had been violated, χ2(2) = 9.874, p = .007, therefore Greenhouse-Geisser corrected tests are reported (ε = .585). The results show that the different lags have a significant effect on the face localization accuracy, F(1.17, 10.53) = 6.502, p = .024.
Table 6
Means face localization accuracy and Standard deviations for the different lags.
Lag 2 Lag 3 Lag 8
Mean SD .680 .123 .752 .142 .799 .114
Figure 8. Means face localization accuracy scores for lags 2, 3, and 8.
For the analysis of the distribution of the scores a Repeated Measures-ANOVA with 3 lags and 4 levels of PAS score was used. There was a significant interaction effect F(6, 54) = 8.634, p < .001.
The graphs of the distribution of button presses for each different lag (Figure 9) shows that in each lag PAS score 1 is pressed the most (61%, 56.6%, and 44.8%). The presses for PAS score 2 also increase as the lag does (25.7%, 28.4%, and 8.2%). These percentages indicate a more gradual shift from PAS 1 to PAS 4, which indicated that visual subjective consciousness is being used in the PAS in a gradual fashion.
Backward Masking measured with the PAS
To see if there was any effect of presentation time on the subjective PAS scale, the mean and standard deviation of the continuous scale scores of each lag has been calculated, see Table 7. Figure 10 shows that the PAS scores increase when presentation time increases.
A Repeated Measures ANOVA was executed for the effect of the different lags on the continuous scores. Mauchly’s test for sphericity indicated that the assumption of sphericity had been violated, χ2(2) = 11.017, p = .004, therefore Greenhouse-Geisser corrected tests are reported (ε = .572). The results show that presentation time has a significant effect on the PAS scores, F(1.44, 10.299) = 38.057, p < .001.
Table 7
Means PAS Scores and Standard deviations for the different lags.
10 ms 20 ms 20 ms + 10 ms blank Mean SD 1.356 .445 1.595 .549 2.066 .674
Table 8 shows the mean face localization accuracy and standard deviations for different presentation times. As shown in Figure 11 the presentation times have effect on the face localization accuracy.
A Repeated Measures ANOVA was used to look for the main effects of presentation time on localization accuracy. Mauchly’s test for sphericity showed χ2(2) = .818, p = .664. The results show that presentation times has a significant effect on face localization accuracy, F(2, 18) = 41.568, p < .001.
Table 8
Means face localization accuracy and Standard deviations for presentation time
10 ms 20 ms 20 ms + 10 ms blank Mean SD .673 .114 .794 .143 .900 .108
Figure 11. Means face localization accuracy scores for presentation time
For the analysis of the distribution a Repeated Measures-ANOVA was used. The ANOVA shows that there is a significant effect of the interaction between presentation time and PAS scores F(6, 54) = 17.385, p < .001.
percentage shift from PAS 1 towards PAS 2 over the different presentation times. In the graphs displaying the distributions of the longest presentation time PAS 2 is pressed most, 43,1%.
Figure 12, Distribution of PAS scores for each presentation time in Backward Masking
To test whether the techniques are reporting in the same manner a Repeated Measure-ANOVA was used. An significant interaction effect between technique and presentation time/lag on PAS score was found F(6, 54) = 4.650, p = .001.
This effect indicated that there is a significant difference between the two techniques in distribution. As shown in Figures 9 and 12 the distributions of the Attentional Blink Paradigm and Backwards Masking are different, but both still lean toward a gradual shift from PAS 1 towards PAS 4.
General Discussion
In this study different subjective scales and techniques were used and looked at to determine whether subjective visual consciousness is a gradual or an all-or-none phenomenon. Both the Attentional Blink paradigm and Backward Masking paradigm were scored with a Continuous Scale and the PAS scale. As an objective measure a face localization test was used and each paradigm was tested with three different lags/presentation times.
The results show that there was an effect of presentation time/lag on the mean scores on both the Continuous Scale as well as the PAS. Presentation time/lag also had a significant effect on the objective face localization task.
For the techniques measured with the Continuous Scale the outcomes in terms of gradual vs. all-or-none differed. In the Attentional Blink task there was a more gradual distribution of the scores, even though it was not perfectly gradual; there was still a steep peak for the first bin. The Backward Masking Paradigm outcomes showed a more all-or-none shift.
The PAS scores in the Attentional Blink Paradigm and Backward Masking Paradigm both showed a more gradual distribution. Even though both the distributions leaned more to a gradual view, they weren’t alike in all aspects. In the Attentional Blink PAS 1 were presses more often for each lag in comparison with the Backward Masking Paradigm; this was also the case for the Continuous Scores.
The first hypothesis discussed in this paper is that all the scales should have the same outcome, either gradual or all-or-none. From the results of the different distribution of each technique/scale the conclusion can be made that this hypothesis cannot be accepted. While both the experiments with PAS scores had a gradual outcome as well as the Attentional Blink paradigm measured with the Continuous Scale, was the outcome of the Backward Masking Paradigm measured with the Continuous Scale more all-or-none.
The second hypothesis drawn states that within the two different techniques the outcomes of the scales should not differ. This means that for the Attentional Blink the outcome of the PAS and Continuous Scale are telling the same story; the same counts for the Backward Masking Paradigm. This hypothesis can be accepted for the Attentional Blink, in which both the outcome of the Continuous Scale and the PAS is gradual. For the Backward Masking there is a different in outcome for the two scales.
The different outcomes for the scales and techniques could be attributed to different factors. The first factor is that the different techniques, Attentional Blink and Backward Masking, have a different difficulty level. From the means of the scores of the two techniques it is clear that the Attentional Blink subjective scores are significantly lower for both the Continuous Score and PAS. When the techniques are not equal in difficulty it is hard to say if you are measuring visual subjective consciousness on the same level, this makes the two
techniques hard to compare within and between the scales.
A second reason for the different outcomes may be the anchoring effect. People have the tendency to choose close by an already given answer. For the continuous scale the only given are the two extreme ends.
Therefor the anchoring effect causes the data to become dichotomous, because subjects have the tendency to choose for the most extreme options. This could have been the case for the Backward Masking experiment measured with the Continuous Scale. An argument against this theory is that if the Continuous Scale encourages subject to choose the extremes this should also be the case in the Attentional Blink Paradigm. However, this
argument can be revoked because in the study of Sergent and Dehaene (2004) they found an all-or-none
distribution for the Attentional Blink Paradigm measured with the Continuous Scale and a gradual distribution for the Backward Masking task also measures with the Continuous Scale and used this same argument.
interpreted. For the PAS scale there are 4 possible peaks in the distribution. It is unclear whether you can speak only of an all-or-none distribution when only the first and fourth peaks are present or that you can also have an all-or-none effect when for example the first and second peaks are present. These different interpretations of the PAS scales depend on how you will interpret the term ‘consciousness’. Therefor it is hard to say whether you can actually use the PAS scale to test for an all-or-none distribution.
A fourth explanation for all the different results could be that it is still unclear whether these different kind of subjective measures and techniques are a valid measure for subjective visual perception. It could be the case that one or both of the measures and technique are testing some other process than consciousness
perception.
Three of the four outcomes indicated a gradual shift between unconscious and conscious perception, because of this it can be stated with caution that it is more likely to be a gradual phenomenon. But given the small sample this can only be looked at as an exploratory research without hard evidence. To further analyze these different outcomes and possible shortcomings there should be done more research on this particular topic. In further research it should be taken into account that the difficulty levels of the techniques may differ and this should be corrected for. There should also be a correction for the anchoring effect that can occur with the Continuous Scale.
Literature
Breitmeyer, B. G., & Ogmen, H. (2000). Recent models and findings in visual backward masking: A comparison, review, and update. Perception & psychophysics, 62(8), 1572-1595.
Dehaene, S. (2014). Consciousness and the Brain: Deciphering How the Brain Codes Our Thoughts. New York, NY: Viking Penguin.
Del Cul, A., Baillet, S., & Dehaene, S. (2007). Brain dynamics underlying the nonlinear threshold for access to consciousness. PLoS Biol, 5(10), e260.
Jiang, Y., Costello, P., & He, S. (2007). Processing of invisible stimuli: Advantage of upright faces and recognizable words in overcoming interocular suppression. Psychological science, 18(4), 349-355.
Nieuwenhuis, S., & de Kleijn, R. (2011). Consciousness of targets during the attentional blink: a gradual or all-or-none dimension?. Attention, Perception, & Psychophysics, 73(2), 364-373.
Overgaard, M., Rote, J., Mouridsen, K., & Ramsøy, T. Z. (2006). Is conscious perception gradual or dichotomous? A comparison of report methodologies during a visual task. Consciousness and cognition, 15(4), 700-708.
Persaud, N., McLeod, P., & Cowey, A. (2007). Post-decision wagering objectively measures awareness. Nature neuroscience, 10(2), 257-261.
Ramsøy, T. Z., & Overgaard, M. (2004). Introspection and subliminal perception. Phenomenology and the Cognitive Sciences, 3(1), 1-23.
Sandberg, K., Bibby, B. M., Timmermans, B., Cleeremans, A., & Overgaard, M. (2011). Measuring
consciousness: task accuracy and awareness as sigmoid functions of stimulus duration. Consciousness and cognition, 20(4), 1659-1675.
Sandberg, K., Timmermans, B., Overgaard, M., & Cleeremans, A. (2010). Measuring consciousness: is one measure better than the other?. Consciousness and cognition, 19(4), 1069-1078.
Schurger, A., & Sher, S. (2008). Awareness, loss aversion, and post-decision wagering. Trends in cognitive sciences, 12(6), 209-210.
Sergent, C., & Dehaene, S. (2004). Is consciousness a gradual phenomenon? Evidence for an all-or-none bifurcation during the attentional blink. Psychological science, 15(11), 720-728.
Seth, A. K. (2008). Post-decision wagering measures metacognitive content, not sensory consciousness. Consciousness and cognition, 17(3), 981-983.
Seth, A. K., Dienes, Z., Cleeremans, A., Overgaard, M., & Pessoa, L. (2008). Measuring consciousness: relating behavioural and neurophysiological approaches. Trends in cognitive sciences, 12(8), 314-321.
Stein, T., Hebart, M. N., & Sterzer, P. (2011). Breaking continuous flash suppression: a new measure of unconscious processing during interocular suppression. Frontiers in human neuroscience, 5(10.3389). Szczepanowski, R., Traczyk, J., Wierzchoń, M., & Cleeremans, A. (2013). The perception of visual emotion:
comparing different measures of awareness. Consciousness and cognition, 22(1), 212-220. Wierzchoń, M., Asanowicz, D., Paulewicz, B., & Cleeremans, A. (2012). Subjective measures of
