How sequence learning creates explicit knowledge: the role of response–stimulus interval

Online Anticipatory Cues During Practice Disrupt Intentional and Incidental Sequence Learning

In the Serial Reaction Time Task, participants respond to several stimuli usually being unaware that the stimuli follow a predefined sequence while still learning the sequence. In the present study, we aimed to clearly separate explicit intentional learning from implicit incidental learning by either informing participants about all details of the sequence or not informing participants about the existence of the sequence. Further, we explored the influence of anticipatory cues during practice while anticipatory cues were either presented (extrinsically triggered anticipation) or not presented (self-reliant intrinsic anticipation). Participants were tested before and after practice in the Practice Sequence and a Control Sequence. To test automatization, tests were performed in Single-Task and Dual-Task Blocks. Results showed that after learning with explicit instructions, participants memorized the sequence more deeply and executed the sequence faster than after learning without explicit instructions. Further, by learning with anticipatory cues, participants memorized the sequence less deeply and executed the sequence slower than by learning without anticipatory cues. Unexpectedly, automatization was sequence-unspecific and independent of the practice conditions. In conclusion, detailed explicit prior information about the sequence facilitates sequence learning while anticipatory online cues during practice hamper sequence learning.

The effect of response-to-stimulus interval on children's implicit sequence learning

The current study examined, for the first time in a developmental perspective, the effect of response-to-stimulus interval (RSI) in incidental sequence learning (SL). Children aged 4, 7, and 10 years performed a serial reaction time (SRT) task in which the RSI was systematically manipulated (0, 250, 500, or 750 ms). SL (difference in reaction times between fixed and random blocks) was not observed for the youngest children whatever the RSI condition, whereas the 7-year-olds learned the sequence only in the 250-ms RSI condition and the 10-year-olds exhibited SL in all temporal conditions except the 500-ms RSI condition. Finally, the results suggest that conscious awareness of the sequence emerges only in older children faced with the 500- and 750-ms RSI conditions. The discussion questions the robustness of implicit learning processes in the light of individual and contextual factors.

Regular rhythmic and audio-visual stimulations enhance procedural learning of a perceptual-motor sequence in healthy adults: A pilot study

Procedural learning is essential for the effortless execution of many everyday life activities. However, little is known about the conditions influencing the acquisition of procedural skills. The literature suggests that sensory environment may influence the acquisition of perceptual-motor sequences, as tested by a Serial Reaction Time Task. In the current study, we investigated the effects of auditory stimulations on procedural learning of a visuo-motor sequence. Given that the literature shows that regular rhythmic auditory rhythm and multisensory stimulations improve motor speed, we expected to improve procedural learning (reaction times and errors) with repeated practice with auditory stimulations presented either simultaneously with visual stimulations or with a regular tempo, compared to control conditions (e.g., with irregular tempo). Our results suggest that both congruent audio-visual stimulations and regular rhythmic auditory stimulations promote procedural perceptual-motor learning. On the contrary, auditory stimulations with irregular or very quick tempo alter learning. We discuss how regular rhythmic multisensory stimulations may improve procedural learning with respect of a multisensory rhythmic integration process.

The gradual subjective consciousness fluctuation in implicit sequence learning and its relevant brain activity

Existing studies have investigated gradual subjective consciousnesses, guess, intuition, fluency, rule, and memory, and their fluctuation behavioral characteristics in implicit learning, but they did not investigate or elucidate the underlying brain mechanisms. Therefore, the current study asked participants to report subjective consciousnesses in each trial of inclusion and exclusion tasks after implicit sequence learning and used the eyes-closed and eyes-opened resting-states' fMRI to examine the relevant brain areas of the five gradual subjective consciousnesses and their fluctuation. The results showed that: (1) There were many relevant resting-state brain areas of the five gradual subjective consciousnesses to reveal their brain mechanisms. In the eyes-closed and eyes-opened resting states, as the participants' consciousness level was gradually increasing from guess to intuition, to fluency, to rule, and to memory, the positively-relevant brain areas correspondingly changed from somatic motor to a mixture of somatic motor, consciousness, emotion feeling, and implicit learning; and then to a mixture of visual, somatic motor, and consciousness; and then to a mixture of visual, somatic motor, and consciousness; and then to a mixture of visual, somatic motor, and consciousness. The negatively-relevant brain areas correspondingly changed from a mixture of visual, consciousness, somatic sensory, and implicit learning to a mixture of visual, somatic motor, somatic sensory, and other consciousness; and then to memory; and then to a mixture of other somatic motors; and then to a mixture of other consciousness and other somatic motors. However, in the amplitude of low frequency fluctuations (ALFFs)-difference, the relative directions of the guess and intuition were almost opposite to those in the eyes-closed and eyes-opened resting states. But the relative directions of the fluency, rule, and memory were consistent with those in the eyes-closed and eyes-opened resting states. (2) There were significant gradual subjective consciousness fluctuations, including the gradual subjective consciousness fluctuation-all M and SD. There were many relevant resting-state brain areas of gradual subjective consciousness fluctuations to reveal their brain mechanisms. The gradual subjective consciousness fluctuation M was positively related to Calcarine_R, Lingual_R, Lingual_R, Temporal_Pole_Mid_L, ParaHippocampal_L, Vermis_1_2, and Vermis_7; but was negatively related to Calcarine_R. The gradual subjective consciousness fluctuation-all SD was positively related to Parietal_Inf_L, Thalamus_L, Temporal_Mid_L, Vermis_9, Parietal_Inf_L, and Thalamus_L and Thalamus_R; but was negatively related to Rolandic_Oper_R, Rolandic_Oper_R, Insula_L, Insula_R, Cingulum_Post_L, and Temporal_Mid_L. The detailed function of the relevant brain areas of consciousness fluctuations needs further investigation. (3) ALFFs in eyes-closed and eyes-opened resting states and their ALFFs-difference could differently predict the five gradual subjective consciousnesses and their fluctuations, indicating that using the two resting states was necessary, and the ALFFs-difference was a new quantitative sensitivity index of the gradual subjective consciousnesses and their fluctuations.

The ecological validity of MET was favourable in sitting implicit sequence learning consciousness by eyes closed and eyes open resting states fMRI

The current study made participants sit to complete both the implicit sequence learning and the inclusion/exclusion tasks with the latter just after the former, and used eyes-closed and eyes-open resting states fMRI and their difference to test the ecological validity of the mutually exclusive theory (MET) in implicit-sequence-learning consciousness. (1) The behavioral and neuroimaging data did not support the process dissociation procedure, but did fit well with the MET. The correct inclusion-task response and the incorrect exclusion-task response were mutually exclusive with each other. The relevant brain areas of the two responses were either different or opposite in the eyes-closed and eyes-open resting-states and their difference. (2) ALFFs in eyes-closed and eyes-open resting-states and their difference were diversely related to the four MET knowledge in implicit sequence learning. The relevant brain areas of the four MET knowledge in the eyes-closed and eyes-open resting-state were the cerebral cortex responsible for vision, attention, cognitive control and consciousness, which could be called the upper consciousness network, and there were more relevant brain areas in the eyes-open resting-state than in the eye-closed resting-state.The relevant brain areas in ALFFs-difference were the subcortical nucleus responsible for sensory awareness, memory and implicit sequence learning, which could be called the lower consciousness network. ALFFs-difference could predict the four MET knowledge as a quantitative transition sensitivity index from internal feeling to external stimulus. (3) The relevant resting-state brain areas of the four MET knowledge were either different (for most brain areas, if some brain areas were related to one MET knowledge, they were not related to the other three MET knowledge) or opposite (for some brain areas, if some brain areas were positively related to one MET knowledge, they were negatively related to other MET knowledge). With the participants' control/consciousness level increasing from no-acquisition to controllable knowledge step by step, the positively relevant resting-state brain areas of the four MET knowledge changed from some consciousness network and the motor network, to some consciousness network and the implicit learning network, and then to some consciousness network; and the negatively relevant resting-state brain areas of the four MET knowledge changed from some consciousness network and visual perception network, to some consciousness network, then to some consciousness network and the motor network, and then to some consciousness network, the implicit learning network, and the motor network. In conclusion, the current study found the ecological validity of the MET was good in sitting posture and eyes-closed and eyes-open resting-states, ALFFs in eyes-closed and eyes-open resting-states and their difference could predict the four MET knowledge diversely, and the four MET knowledge had different or opposite relevant resting-state brain areas.

Testing the Process Dissociation Procedure by Behavioral and Neuroimaging Data: The Establishment of the Mutually Exclusive Theory and the Improved PDP

The process dissociation procedure (PDP) of implicit sequence learning states that the correct inclusion-task response contains the incorrect exclusion-task response. However, there has been no research to test the hypothesis. The current study used a single variable (Stimulus Onset Asynchrony SOA: 850 ms vs. 1350 ms) between-subjects design, with pre-task resting-state fMRI, to test and improve the classical PDP to the mutually exclusive theory (MET). (1) Behavioral data and neuroimaging data demonstrated that the classical PDP has not been validated. In the SOA = 850 ms group, the correct inclusion-task response was at chance, but the incorrect exclusion-task response occurred greater than chance. In the SOA = 850 ms group, the two responses were not correlated, but in the SOA = 1,350 ms group and putting the two groups together, the two responses were in contrast to each other. In each group, brain areas whose amplitude of low frequency fluctuations (ALFFs) in the resting-state related to the two responses were either completely different or opposite to one another. However, the results were perfectly consistent with the MET proposed by the present study which suggests that the correct inclusion-task response is equal to the correct exclusion-task response is equal to C + A₁, and the incorrect exclusion-task response is equal to A₂. C denotes the controlled response and A₁ and A₂ denote two different automatic responses. (2) The improved PDP was proposed to categorize the 12 kinds of triplets as delineating four knowledge types, namely non-acquisition of knowledge, uncontrollable knowledge, half-controllable knowledge, and controllable knowledge with the MET. ALFFs in the resting-state could predict the four knowledge types of the improved PDP among two groups. The participants’ control of the four knowledge types (degree of consciousness) gradually improved. Correspondingly, the brain areas in the resting-state positively related to the four knowledge types, gradually changed from the sensory and motor network to the somatic sensorimotor network, and then to the implicit learning network, and then to the consciousness network. The brain areas in the resting-state negatively related to the four knowledge types gradually changed from the consciousness network to the sensory and motor network. As SOA increased, the brain areas associated with almost all the four knowledge types changed. (3) The inhomogeneous hypothesis of the MET is best suited to interpret behavioral and neuroimaging data; it states that the same components among the four knowledge types are not homogeneous, and the same knowledge types are not homogeneous between the two SOA groups.

Explicit instructions facilitate performance of OCD participants but impair performance of non-OCD participants on a serial reaction time task

Previous studies have shown that individuals diagnosed with OCD tend to rely on explicit processing while performing implicit learning tasks. We sought to investigate whether individuals with OCD are capable of implicit learning, but would demonstrate improved performance when explicit processing strategies are enhanced. Twenty-four participants with OCD and 24 non-psychiatric control (NPC) participants performed an implicit learning task in which they responded to a single target stimulus that successively appears at one of four locations according to an underlying sequence. We manipulated the learning strategy by informing half of the participants that the target stimulus location was determined by an underlying sequence, which they should identify (intentional learning). The other half of the participants was not informed of the existence of the underlying sequence, and was expected to learn the sequence implicitly (standard learning). We predicted that OCD participants will exhibit inferior performance compared to NPC participants in the standard learning condition, and that intentional learning instructions would impair the performance of NPC participants, but enhance the performance of OCD participants. The results supported these predictions and suggest that individuals with OCD prefer controlled to automatic processing. We discuss the implications of this conclusion to our understanding of OCD.

fMRI Investigation on Gradual Change of Awareness States in Implicit Sequence Learning

Awareness of implicit knowledge is a changing process. Previous studies have examined brain activation patterns corresponding to the start and end stages of implicit learning, but failed to reveal the gradual changing course of awareness in implicit learning. The present study explored brain activation changes corresponding to different awareness states elicited by two different stimulus onset asynchrony (SOA, 850 ms and 1350 ms) over the whole course of implicit sequence learning (i.e., divided into three phases), by using a process dissociation procedure (PDP) paradigm and the technique of functional magnetic resonance imaging (fMRI). In the results, it was found that the 850 ms SOA elicited primarily an awareness state of unconsciousness, under which the frontal lobe was significantly activated during the early phase of implicit sequence learning, with its activation levels correlated positively to consciousness levels. In contrast, the 1350 ms SOA triggered predominantly an awareness state of consciousness, under which the activation levels of the inferior parietal lobule correlated positively to consciousness levels during the middle phase, and positively to consciousness levels as well as negatively to unconsciousness levels during the late phase of implicit sequence learning. Overall, the frontal lobe and inferior parietal lobule were found to play critical roles in mediating awareness states over the course of implicit sequence learning.

Action-effects enhance explicit sequential learning

Different studies have shown that action-effect associations seem to enhance implicit learning of motor sequences. In a recent study (Haider et al., Conscious Cognit 26:145-161, 2014), we found indications that action-effect learning might play a special role in acquiring explicit knowledge within an implicit learning situation. The current study aims at directly manipulating the action-effect contingencies in a Serial Reaction Time Task and examining its impact on explicit sequence knowledge. For this purpose, we created a situation in which the participants' responses led to a melodic tone sequence. For one group, these effect tones were contingently bound to the sequential responses and immediately followed the key press; for the second group, the tones were delayed by 400 ms. For a third group, the tones also followed the response immediately and resulted in the same melody but were not contingently bound to the responses. A fourth control group received no effect tones at all. Only the group that experienced contingent effect tones that directly followed the response showed an increase in explicit sequence knowledge. The results are discussed in terms of the multi-modal structure of action-effect associations and the ideomotor principle of action control.

Age and time effects on implicit and explicit learning

BACKGROUND/STUDY CONTEXT: It has been proposed that effects of aging are more pronounced for explicit than for implicit motor learning. The authors evaluated this claim by comparing the efficacy of explicit and implicit learning of a movement sequence in young and older adults, and by testing the resilience against fatigue and secondary tasking after learning. It was also examined whether explicit learning in older adults can be promoted by alleviating time constraints during learning.

METHODS

The alternating serial reaction time task (ASRTT) was used. Experiment 1 compared the benefits of receiving full instructions about the stimulus sequence relative to receiving no instructions in young (20-25 years) and older (50-65 years) adults during retention and during transfer to fatigue and secondary task conditions. Experiment 2 alleviated time constraints during the initial bouts of practice with full instructions.

RESULTS

Experiment 1 indicated that the older adults learned on the ASRTT and achieved similar performance as young adults when no instructions were given. In contrast to the young adults, learning was not superior in older adults who received full instructions compared with those who did not. Experiment 2 indicated that alleviating time constraints allowed some of the older adults to gain from instruction but only under relatively low time constraints, but there was no retention with rigorous time constraints.

CONCLUSION

Explicit learning, but not implicit learning, declines in older adults. This is partly due to older adults difficulties to apply explicit knowledge. Less rigorous time constraints can help to ameliorate some of these difficulties and may induce levels of explicit learning in older adults that will result in superior performance compared with implicit learning. Implicit learning did occur under time constraints that prevented explicit learning.