We have included in this study four different groups of participants: two composed of adolescents participating in Dialectical Behavior Therapy (DBT) skills training and two others consisting of their respective parents, also participating in DBT skills training. The demography of the groups is presented in Table 1.

The adolescents followed two slow-open DBT skills training groups of 2-hour sessions every week. One of the participants was excluded from the study because she did not participate in any group sessions. The group of parents was a slow open group of 10 DBT sessions of 1.5 hours each, also weekly.

The study started after the Basurto University Hospital Ethics Committee (Bilbao, Spain) gave its approval in adherence to the Helsinki Declaration for research with human subjects. All participants gave written informed consent after receiving oral and written information about the experiment, and specifically for adolescents, both the participant and their parents or legal tutor signed

informed consents. All participant data were coded to be completely anonymous, including for the researchers analyzing the data.

The skills taught in these groups are mindfulness, distress tolerance, regulation of emotions, interpersonal effectiveness, and walking the middle path. The group for adolescents consists of 16 2-hour sessions weekly. And the group for parents consists of 10 sessions of 1.5 hours duration, also weekly.

The groups included in this study were “slow-open,” which means that participants could enter and leave the group during the training. Two therapists ran each group.

Before joining the group, all participants (adolescents and parents) attended an interview for evaluation and information on the group’s methodology and the research. During that interview, participants filled the informed consent and the questionnaire number “zero” (with one exception). In addition, sociodemographic data on adolescents were collected (see Table 1).

An identification code was assigned to each participant to keep his identity anonymous for the research. We have extensively described the general setting of this experiment in previous publications (Fernandez-Rivas et al., 2020; Trojaola-Zapirain et al., 2014, 2015, 2016, 2019). It will be enough to recall that we have used a questionnaire composed of 50 pairs of figures for this work. Participants were asked to select one picture from each pair and to complete the questionnaire in three minutes. The pictures’ choice aims to minimize the socio-cultural bias introduced by a word questionnaire (Zanello et al., 2004). The figures in each pair were always the same, but the pairs were randomly reshuffled at each test repetition to minimize mnemonic effects. Figure 1 reports a sample page from the questionnaire with fictional picture choices.

We will name the four groups PG1 (first parent group), PG2 (second parent group), YG1 (first adolescent group), and YG2 (second adolescent group). Not all the parents participated in the groups. The duration of the four training

courses is, respectively, 43, 31, 43, and 27 sessions.

The participation has been somewhat uneven to the different groups (see Figures 2-4). The presence in the group of adolescents was less regular after the 5^th session.

The most frequently chosen picture in each pair during the “zero” test (the questionnaire administered before the first meeting) is indicated as picture A (A_i, i = 1.50), while the other picture is designed as B (B_i, i = 1.50). We have computed frequency tables for each pair of images and each session for the four groups. Because the present work is focused on the influence of the group unconscious on the measured effects—i.e., the answers to the questionnaire—we consider only the proportion of the participants choosing picture A or B for each of the 50 questions, irrespectively of how the individual participant’s choice evolved.

Given the very uneven participation, we decided not to correct for data that were completely missing. Some of the sheets were incorrectly marked, and those we corrected with an LOCF (Last Observation Carry Forward, (Hamer & Simpson, 2009)) algorithm, using the same answer from the previous session. If this was not possible, the session before was used, and so on. There was no instance of invalid questionnaires for both the zero and first sessions. The total number of corrections is reported in Table 2.

We report the evolution of the A’s (most chosen picture in questionnaire zero) for the four groups during the training. We have compared the percentage of the “A” picture choice for the 50 questions in the successive sessions. For this, we have performed a Friedmann test on the whole set of the × 50 data, and then, if significant—which has always been the case—we have performed a post-hoc Conover test (Conover, 1999; Conover & Iman, 1979) between

successive sessions. All the data analysis has been done with the R software (R Core Team, 2020).

We show the statistically significant difference in the choice of A’s between sessions, including the initial questionnaire (see Figure 5 and Table 3).

We consider now how the choice of A and B has evolved comparing the changes of choice of the participants. For each couple of successive sessions, we count the number of participants whose choice has changed from A to B and the number of those whose choice has changed from B to A. We indicate this transition with the number of the second session, i.e., we indicate with n the transition (A→B o0r B→A) between session n−1 and session n. To avoid repetitions, we will refer to this explanation as the notation explained in the section “Evolution of the transition between questionnaires.”

We first consider the number of changes from A→B. The statistical differences we found for the four groups are reported in Table 4 and Figure 6.

Now we consider the same statistics but for the transitions B→A for the four groups. The results are shown in Figure 7 and Table 5.

We now investigate whether the individual changes of choice A→B and B→A are simple statistical fluctuations in people’s preference or whether there is a statistically significant difference. This may indicate that the change of choice of the group taken as an ensemble is not casual and that there is a significant movement either toward the initial choice or away from it. As before, we indicate with n the change from A→B or B→A between session n−1 and session n. We will call this the non-casuality plot.

We found statistically significant differences, considering the four groups, with the notation explained in the session “Evolution of the transitions between questionnaires,” as shown in Figure 8 and Table 6.

We can combine the transitions A→B and B→A into two derived quantities that help to interpret the evolution of the groups’ choice of pictures. We call the quantity A→B+B→A the “flux” as it expresses the total “activity” of the group in changing the initial choice of picture. We introduce a second quantity, B→A-A→B, that we call “focus” as it indicates the tendency to confirm, if greater than 0, or move away from, if smaller than 0, the initial choice of A.

As in the previous section, we consider at first a Friedmann test of the whole 50 x matrix for the focus for the four groups. All these tests indicate a significant deviation from purely random fluctuations. Following this, we compare each session with the following one using the Conover post-hoc test for each group. The results can be seen in Figure 9 and Table 7.

We now consider the focus, i.e., the difference of the changes of choice B→A-A→B that gives us the tendency to converge toward the initial choice—if positive—or to move away from it—if negative. The results can be seen in Figure 10 and Table 8.

In Figure 11, we show the evolution of the percentage of A’s choice in chronological order for the four groups.

In Figure 12, we show the statistically significant differences between sessions for the percentage of A’s choice and the percentage of A→B and B→A transitions in chronological order for the four groups. For the latter two, we use the notation explained in the section “Evolution of the transitions between questionnaires.”

Considering the choice of A, we show the statistically significant comparisons between consecutive sessions, including the questionnaire zero completed before the beginning of the training. For the transition A→B and B→A, we use the

notation explained in the section “Evolution of the transitions between questionnaires.”

In Figure 13, we show the statistically significant differences between sessions for the percentages of focus (B→A-A→B), flux (A→B+B→A), and non-casuality (A→B vs. B→A) transitions in chronological order for the four groups. This allows us to determine whether there is a correlation between non-casuality, flux, and focus during the four groups’ chronological evolution. We remember that we have explained the notation used for the questionnaire number in the section “Evolution of the transitions between questionnaires.”

We believe that the best way to study the unconscious is to observe its amplification in group dynamics. As it happens in quantum mechanics, the problem of measurement is central also in psychophysics because the unconscious is, by definition, unknowledgeable and not only because it is “unconscious” but also

because the “detector” is the cognitive part of the individual, which is deeply influenced by, and indeed built upon, the unconscious.

This analogy relies on the observation that group dynamics, as described by Bion’s “basic assumptions,” is similar to individual dynamics, in particular in the crucial aspect of the analogy of the mourning process in the individual with the loss of the ideal leader in the group, inspired by the Oedipean constellation.

Considering the simultaneous presence of potentially connected individuals, we have made the hypothesis that the entanglement effects could be more pronounced in the case of a group setting. We call this “quantum amplification.” As said in the Introduction, Jung (Jung & McGuire, 1925) was aware that the individual psyche is tributary of a layered unconscious, where the layers are the “collective unconscious” of larger and larger groups, such as the family and society.

Jung also speaks of amplification (Jung & Hull, 1911). For Jung, the “amplification” is the extension and the deepening of a dream-like image by means of associations centered on the dream theme and associations based upon social studies and the history of symbols (mythology, mysticism, folklore, religion, ethnology, art, etc.). Thanks to this, the dream becomes accessible to interpretation (Jung, 1962). If we consider the analogy conveyed by the term “amplification” in the two contexts, unconscious mental processes like dreams can be regarded as “microscopic” quantum processes, becoming accessible to conscience only via an amplification and measurement process, in this case, operated by consciousness. This is one more example of the interesting parallels that we can draw between quantum physics and psychodynamics.

In the case presented in this article, we can suppose that the group situation generated by the DBT skills training is influenced not only by the individual and family situations but also by the environment, such as the large social group.

In quantum physics, a microscopic process is “amplified” by the “observer” to the macroscopic level, allowing the measurement to take place. Thus, only after amplification, we can observe a microscopic quantum process as a physical phenomenon via an act of measurement. However, the irreversibility of such an act is still an open argument of discussion.

The perspective of this study is to observe if the group dynamics—particularly in a slow-open setting—is also affected by the environment (as in quantum mechanics).

In the present experiment, we noted some differences between parents and adolescents—for example, in the numbers of sessions, similar to what we have observed in one of our previous work of ours (Fernandez-Rivas et al., 2020)—but the comparison between parents and adolescents groups will be the subject of a future study.

The most striking aspect of our data is that the answers to the questionnaire zero, administered before the training, are very skewed, with values of A’s percentages respectively of 70.6% for PG1, 70.5% for PG2, 63.7% for YG1, 64.7% for YG2. This is similar to what we found in a previous study of DBT closed groups (Fernandez-Rivas et al., 2020; Trojaola-Zapirain et al., 2019) at the end of the training.

We observed an increase to 74.7% for PG1 and to 82% for PG2, and a reduction to 52% for YG1 and 56% for YG2.

The trends in the changes of the A’s answers are different in the four groups, and it is interesting to note that both groups of parents, PG1 and PG2, show an increase in the choice of the initial preferred image (+4.1% and +11.5%) while both groups of young YG1 and YG2 show a decrease (−11.7% and −5.6%).

The changes A→B and B→A in each of the four groups are similar, and this is confirmed that there are few non-casuality significant differences. We recall that we indicate as non-casuality the fact that the changes A→B and from B→A are not in equilibrium, i.e., the percentage of changes of the 50 questions are not statistically comparable for the two directions.

The flux, being the sum of A→B and B→A, presents statistically significant values almost at the same moment as A→B and B→A. On the other hand, the statistically significant values for the focus are less than in A→B, B→A, and flux.

The interpretation of this observation is not simple: if we imagine that the originally preferred answer—the A answer of the questionnaire zero—is the “personal” answer independent from the influence of the group, we can suppose that the return to this choice signals a movement “against” the “group continuum” and toward loyalty to the familiar and social clan that we can define as “normal clanity.”

The PG1 group shows many significant variations of the flux (but also of A→B and B→A) between different sessions, more than in the other groups. We note that focus and non-casuality present significative differences in the same phase of the training, i.e., before the 20^th session and at the end of the training (except for PG1). This is consistent with the “canonical” group evolution. At the beginning of the training, the group experiences the groupal disillusion. At the end, the group reacts against and mourns the conclusion of the experience, and participants tend to return to their normal clanity to preserve themselves from the group’s disillusion/demise.

Because the groups had a training and therapeutical goal, the fact that some groups took place simultaneously was casual and independent from the research objectives. Nevertheless, this overlap in time of the different groups allows a comparison of their evolution in time. Probably because of the slow-open structure of the groups, we observed some coincidence in their evolution which could depend more on external factors than on the group’s internal evolution.

If we observe the group evolution in calendar time rather than in the number of sessions, we see the same coincidences in the statistically significant values of some of the parameters considered (A, A→B, B→A, flux), which seem to obey not only the internal evolution of the groups but also other possible external factors. This coincidence was not present for non-casuality and focus.

We underline that if the flux is the expression of changes from A to B plus B to A, the non-casuality indicates the statistical difference of the transitions A→B and B→A, and the focus, defined as B→A minus A→B, is the expression of the tendency to revert to the initial choice of pictures before the beginning of the training.

The interaction with the external environment in slow-open groups is probably experienced by the participants in an interactive mode with important clan loyalties: the high “orientation” (A’s choices before the constitution of the group) is “recovered” and in a group even exceeded during the sessions and the answers are less casual (less 50%) than in a state of “group continuum.” We note that the participants remain in the group for more or less than 10 - 12 sessions. Therefore, we can suppose that the slow-open group promotes a “clan continuum” due more to the external environment than a group continuum.