Effects of cortical polarization on mental arithmetic.

A biofeedback procedure was used to influence participants' cortical polarity before the presentation of single digit multiplication problems. To ensure that participants could solve the problems by direct memory retrieval of arithmetic facts, only nine different multiplication problems were used, and participants received extensive pre-experimental practice on these problems. After biofeedback training, cortical positivity before problem presentation was associated with faster response initiation times, a measure of mental calculation time, for correctly solved problems and with more multiplication errors than cortical negativity. Response execution time, a measure of motor speed, was not affected by cortical polarity. The shift in positivity from the amplitude of the slow cortical potential (SCP) before problem presentation to the amplitude of the positive slow wave (pSW) after problem presentation was less in positivity-required compared to negativity-required trials, mainly because of the manipulation of SCP starting points by biofeedback. We assume that cortical positivity is associated with direct retrieval of arithmetic facts, more precisely with the inhibition of incorrect answers. Cortical positivity before task presentation results in a reduced positive shift, a sign of less inhibition before answer production. The effect of less inhibition during positivity-required trials is to produce faster responses and more errors. The opposite might to be true of cortical negativity.

Learned changes of brain states alter cognitive processing in humans.

Humans can learn to intentionally control their brain states based on information about their own electrocortical activity. Using an operant conditioning technique, twelve healthy volunteers were trained to shift their slow cortical potentials recorded from left-hemispheric language cortices in the positive versus negative direction. After training, six subjects who achieved reliable control of left-hemispheric brain responses showed substantial modification of word processing. During conditioned negative shifts of cortical potentials (activation condition), responses to words were substantially speeded, whereas lexical decisions were slower during positive shifts of slow cortical potentials (inhibition condition). No comparable difference was seen in trained participants who failed to achieve control over slow cortical potentials. Additional data suggest that the effect was not related to perception, attention, or motor processes. Thus, operant conditioning can produce focal cortical activity dynamics and thereby modify specific higher cortical processes such as access to words. This finding may open new perspectives on neuropsychological rehabilitation based on operant conditioning of brain responses.

Semantic or lexico-syntactic factors: what determines word-class specific activity in the human brain?

Words from different classes have been found to activate different brain areas. However, it is unclear whether grammatical word properties, for example their being part of different lexical categories (e.g. nouns vs. verbs) or semantic features of the words (e.g. that they refer to visually perceived entities or to actions) are relevant for eliciting differential brain responses. We tested this by comparing brain potentials elicited by nouns and verbs that varied in their action and visual associations. Naturally spoken word stimuli were from three categories: (1) nouns with strong visual associations; (2) action verbs with strong associations of actions, and (3) nouns with strong action associations. Word-category specific differences became apparent around 500 ms after stimulus onset, approximately 150-200 ms after the average recognition point of the words involved. Brain responses to visual nouns and action verbs differed at central and occipital recording sites. A very similar topographical difference emerged from the comparison of visual vs. action nouns, whereas no significant difference was found between action-related nouns and verbs. These results indicate that grammatical differences alone, e.g. between two lexical classes such as action verbs and action-related nouns, are not sufficient for eliciting differential brain responses. In contrast, semantic differences between items from the same lexical category can be sufficient for changing the topography of cortical processes induced by word stimuli. This is support for associative theories of word processing.

Self-regulation of interhemispheric asymmetry in humans.

Five healthy right-handed subjects learned to control hemispheric asymmetry with biofeedback of the amplitude difference of slow cortical potentials between the left and the right precentral areas. Six training sessions were conducted with subject I, 12 sessions with subjects II and III, and 14 sessions, with subjects IV and V. Performance of four out of five subjects improved continuously as a function of sessions. Towards the end of training, these subjects demonstrated highly significant differentiation between conditions where right versus left precentral negativity was required. In subject V, no improvement was observed after 14 training sessions. The data indicate that most subjects can learn to self-generate fast electroencephalograph (EEG) differences between the left and the right sensorimotor cortical regions.

Frontal interhemispheric asymmetry: self regulation and individual differences in humans.

Sixteen subjects naive to biofeedback learned lateralised interhemispheric control of slow cortical potentials (SCPs) across electrode sites F3-F4 during three sessions of visual electroencephalograph (EEG) biofeedback. Subjects were required to generate slow negativity shifts either towards the left or the right hemisphere in sixty pseudorandomly ordered trials per session. Group 1 (n = 8) were told to use emotional strategies in the task (positive emotions for left hemisphere activation, negative emotion for right hemisphere activation), group 2 received no guidance. Both groups received feedback in the form of an on-screen rocket-ship, initially centrally placed, which rose to indicate an increase in left hemisphere negativity (relative to the right hemisphere) and fell to indicate an increase in right hemisphere negativity (relative to the left hemisphere). A 2 x 3 x 3 x 2 ANOVA (group x session x block x trial) showed no performance differences between the strategy and no strategy groups. Both groups learned to produce correct direction shifts in the final third of each session during both trial types (P < 0.001). The no strategy group showed a particularly strong within session learning effect (P < 0.0037) with poor performance in the early part of the sessions, and strong shifts at the end. Subjects high on withdrawal showed stronger rightward shifts in keeping with right hemisphere involvement in behavioural withdrawal. This is the first demonstration of self regulation of interhemispheric frontal asymmetry.

Self-regulation of slow cortical potentials in epilepsy: a retrial with analysis of influencing factors.

Twenty sessions of biofeedback training were carried out with 12 drug-resistant patients with focal epilepsy who learned to produce either negative or positive shifts of their slow cortical potentials (SCPs) at vertex. Feedback trials were interspersed with transfer trials in which only a discriminative stimulus (signalizing whether positivity or negativity was required) was presented, without feedback signal. Patients were able to differentiate significantly between the conditions of cortical positivity and cortical negativity, with larger differentiation scores being obtained in feedback trials than in transfer trials. The amplitude of positivity generated in the positivity condition increased linearly across sessions both in feedback and in transfer trials. The largest negativity was produced in the 5th session; after this, more transient negativities were generated, whose amplitude decreased towards the end of trial. The mean severity of seizures, estimated as the frequency of seizures weighted by their subjective 'strength', decreased significantly after training as compared to the pre-training phase. The data suggest that (1) patients could learn to achieve a state of cortical disfacilitation and (2) with progressed learning, they became less motivated for (or afraid of) producing considerable negative shifts, since extensive negativity may reflect cortical over-excitation and therefore be associated with early signs of seizures. The inability of producing cortical negativity is however not necessarily a bad predictor.

Operant conditioning of left-hemispheric slow cortical potentials and its effect on word processing.

This study investigated whether language-related cognitive processes can be modified by learned modulation of cortical activity. Study participants received feedback of slow cortical potentials (SCPs) recorded above left-hemispheric language cortices and were reinforced for producing negative and positive shifts upon two different discriminative stimuli. In all subjects who achieved reliable control of left-hemispheric brain responses, substantial modification of word processing was observed. Behavioral modification could be documented in two experiments in which word probes were presented following discriminative stimuli. When negative shifts of the EEG were required, lexical decisions on words were substantially speeded, while they were slowed during positivity conditions. There was no indication for any performance difference between conditions in control subjects who failed to achieve control over SCPs after feedback training. This result was replicated in an experiment using lateralized-tachistoscopic stimulus presentation. Comparisons of word and pseudoword responses in both experiments indicated that behavioral modification was most pronounced for word responses. It was also not seen in a simple reaction time task not involving language materials. This argues against a global effect related to perception, visuo-spatial attention, or motor processes. We conclude that linguistic processes can be influenced by modification of cortical activity due to operant conditioning. In closing, tentative explanations of the present results based on theories of language and attention processes are being discussed.

A new method for self-regulation of slow cortical potentials in a timed paradigm.

A new method of slow cortical potential (SCP) biofeedback is described, in which subjects were presented with a sequence of two alternating tones. Subjects learned to adjust their SCPs with the 4-s rhythm of presented tones by producing directed SCP changes only in certain inter-tone intervals. Specifically, they learned to simultaneously produce two EEG signals: 1) positive or negative SCP shift at vertex, and 2) SCP asymmetry between the right and the left central area. After one training session, 13 healthy participants were able to differentiate significantly between the negativity and the positivity conditions; this differentiation was achieved within less than 300 ms after the discriminative signal, i.e. much faster than in previous studies employing traditional SCP biofeedback technique. However, these participants did not produce a significant hemispheric asymmetry in the first session. In the second experiment, five subjects participated in prolonged training (6 to 17 sessions). Highly significant control of SCP asymmetry over the precentral cortex was attained in four out of five participants. Advantages and disadvantages of the new method as compared with the "classical" SCP biofeedback technique are discussed.

Beyond habituation: long-term repetition effects on visual event-related potentials in epileptic patients.

Sixteen patients with partial epilepsy learned to produce positive or negative slow cortical potential shifts in a biofeedback condition during 20 consecutive training sessions. Visual ERPs to the presentation of the feedback and the discriminative stimulus were recorded at vertex. Regardless of the subjects' task (positivity versus negativity), amplitudes of the P2 (mean peak latency about 225 ms) and P3a (322 ms) components decreased across sessions, resulting in appearance and subsequent enhancement of a negative wave N2 (298 ms) between P2 and P3a. As N2 grew the P2 latency decreased and the P3a latency increased. Additionally, the P3b (472 ms) decreased with repetition, however, it did so slower than P2 and P3a. A comparison between the present data, on the one hand, and those obtained in the ERP habituation paradigm within one session, on the other hand, indicates that some repetition effects cannot be explained by habituation.