Generators of the late cognitive potentials in auditory and visual oddball tasks

https://doi.org/10.1016/S0013-4694(97)00119-3Get rights and content

Abstract

Recordings directly within the brain can establish local evoked potential generation without the ambiguities always associated with extracranial electromagnetic measures. Depth recordings have found that sensory stimuli activate primary cortex and then material-specific encoders. Sensory-specific areas remain active for long periods, but by about 200 ms are joined by activation in widespread brain systems. One system is related to the orientation of attention. It is centered in paralimbic and attentional frontoparietocingular cortex, and associated with the P3a. A second system associated with P3b envelopes cognitive contextual integration. It engages the ventral temporofrontal event-encoding cortices (inferotemporal, perirhinal, and ventrolateral prefrontal), association cortices (superior temporal sulcal and posterior parietal), and the hippocampus. Thus, even in simple tasks, activation is widespread but concentrated in particular multilobar systems. With this information, the late cognitive potentials can be used to monitor the probable location, timing and intensity of brain activation during cognitive tasks.

Introduction

These have the advantages of great temporal resolution and a direct relation to neuronal information-processing. Information is carried between neurons, and is integrated within neurons via current flowing across active brain synapses. In some circumstances, the resulting net extracellular current flow can be recorded on the scalp as the EEG. That is, the EEG is the result of the passive instantaneous electrical propagation from active brain synapses to the scalp recording electrode. When the EEG is averaged with respect to a repeated behavioral event, random background EEG will cancel and only that part of the EEG (termed the EP) related to the behavioral event will remain. Careful examination of EPs across many tasks and subjects has demonstrated that they are composed of a series of components, each defined by its latency, polarity, scalp topography, and behavioral correlates (Halgren, 1990). Successive EP components are related to successive stages in information-processing, from the strictly sensory to the highest integrative levels, termed `endogenous'. Since these EP components are generated by synaptic current flows, they could provide a critical link between cognitive and neural processes. That is, if the intracranial generators of scalp EP components could be identified, then the intensity, onset and duration of activation of specific brain systems could be monitored, without risk or expensive equipment, in normal subjects during cognitive tasks, and functional models for the role of these synapses in generating behavior could be tested. In addition, these functional probes would be of great utility in the basic understanding and clinical evaluation of neurological and psychiatric disorders.

Unfortunately, one cannot unambiguously infer the location of the synapses that generate an EP component (i.e. the `propagating generator') from its scalp topography. This process of estimating cerebral generators from an observed scalp EP topography is known as `solving the inverse problem'. In theory, an infinite number of different generator configurations in the brain could result in the same EP topography at the scalp. Laplacian transforms and spatial deconvolutions can estimate the voltage distribution at the exterior cortical surface with a 2 cm accuracy (see Gevins, 1998, and Koles, 1998). However, only about 30% of the cortex can be monitored in this fashion, and even the locally recorded EPs from electrodes on the cortical surface may be generated in deeper sites.

In the simplest and still most common solution of the inverse problem, the scalp EP distribution is assumed to arise from a single dipole. First, initial values of that dipole's location, orientation and strength are approximated. The propagation of electrical potentials from that dipole to the recording electrodes at the scalp is then calculated analytically by modeling the head as concentric spheres (brain, cerebrospinal fluid, skull and scalp) of differing conductances. The error between the calculated and the measured electrical field patterns is then used to modify the dipole's parameters, and the resulting field is re-calculated. This process is repeated in an iterative manner until the dipole's position, orientation and strength no longer change significantly between iterations.

Using intracranial microstimulation (Cohen et al., 1990; Cuffin et al., 1991; Gharib et al., 1995), or comparison with electrocorticography (Nakasato et al., 1994), the accuracy of such models in localizing an assumed single intracranial generating dipole from extracranial EEG is fairly high, but cognitive EPs do not satisfy this assumption: as will be shown below, cognitive EPs are generated by extended surfaces in multiple brain areas, rather than by a single dipole. Dipole localization methods that take into account the temporal as well as the spatial evolution of the EPs provide an additional constraint, but are still ill-posed (Sherg and VonCramon, 1985).

Synaptic activation not only results in the extracellular current flows that generate the EEG, it also results in intracellular current flows that are the main generators of the MEG. Compared to the EEG, the MEG is very little affected by the type and location of tissue surrounding the generator, and especially that of tissue lying between the generator and the sensor (Hari and Lounasmaa, 1989). However, it is still not possible to unambiguously infer from the topography of the extracranial magnetic field the location of the synapses that generate it. Like the EEG, the MEG will cancel unless there is adequate spatiotemporal synaptic synchrony, and furthermore, radially oriented current flows cannot be detected using MEG. Nonetheless, the MEG is sensitive to different generators than the EEG, and thus it is useful to have both sources of information in order to distinguish between possible generator configurations (e.g. Wood et al., 1985).

These can be used to survey the entire brain for localized changes in glucose metabolism or blood flow that follow neural activity in cognitive tasks. However, the fMRI signal to a single stimulus has a delay of about 2 s and a gradual rise of 3–5 s, and the temporal resolution of PET is at best 45 s. Consequently, the temporal resolutions of PET and fMRI are insufficient to resolve individual cognitive stages.

A linear approach to the inverse problem has been developed that constrains the solution to lie within the cortex and perpendicular to its surface, and then uses a posteriori variance estimates for sources based on sensor spatial covariance (Dale and Sereno, 1993). In addition, the solution can be biased toward areas that were shown to be activated in the same task by fMRI (Nenov et al., 1991; Dale et al., 1995). Modeling studies have suggested, that within limits, this method of integrating fMRI with MEG and EEG data is very promising for arriving at reasonable hypotheses for the spatiotemporal activation pattern of local cortical regions during cognitive tasks. However, it must be emphasized that the result of these calculations is still a hypothesis that needs to be confirmed by direct methods.

Scalp EP recordings after brain lesions in humans have provided important information regarding possible brain generators of scalp EPs (Knight, 1990). Clearly, destruction of a propagating generator would usually be expected to diminish the corresponding scalp EP component. However, it is difficult to evaluate the possible influences of remaining generators (which could show a compensatory increase in strength), of the removal of `canceling' generators (which could actually result in a post-lesional increase in amplitude), and of any skull defect (which alters the distribution of current and thus the scalp EP amplitude). Furthermore, unless synaptic current flows are somewhat synchronous and spatially aligned, they will cancel and remain `occult' to a remote electrode. Finally, the activated synapses within the EP generating structures may project from neurons in another `trigger structure' that in turn relies upon calculations performed elsewhere in some `antecedent structure' (for a more complete discussion of these interpretative issues see Halgren et al., 1986). In order to identify the brain stages of information-processing, occult as well as propagating generators, and trigger as well as antecedent structures must be located. It would be expected that the lesion of a trigger would produce an EP deficit topographically more extensive than the lesion site would suggest, and lesion of an antecedent structure would produce a task-specific EP reduction.

This article focuses on the results from our laboratory using direct intracranial recordings in humans to localize the generators of cognitive EPs.

Section snippets

Methods

The most direct indication that a given structure is a generator of a particular EP component is when recordings from that structure locate large focal polarity-inverting potentials with similar timing and task correlates as that component. Such recordings are possible using electrodes implanted inside the brains of epileptic subjects for the purpose of localizing their epileptic focus prior to its surgical removal (Chauvel et al., 1996). During the time that the patients are waiting for the

Intracerebral potentials to rare target and distractor auditory and visual stimuli

Subjects received an auditory discrimination task with target and non-target rare stimuli (`standard oddball paradigm'). In some cases, the target, distracting and frequent tones were completely balanced across blocks for pitch and volume. Variants included an analogous visual discrimination task, or auditory tasks where the rare target event was the omission of a tone, or the repetition of a tone within a series of alternating tones. In some subjects, the same auditory stimuli were delivered

Organizing principles of late cognitive EPs

Although limited, intracranial recording studies of cognitive EPs permit one to propose some general organizing principles characterizing the human brain's temporospatial pattern of activation during cognitive tasks (Halgren and Marinkovic, 1996). The most striking conclusion is that most areas are activated by a task even if they are not necessary for its performance.

For example, as shown above, simple `oddball' tasks engage widespread neocortical and limbic areas. Given that simple sensory

Conclusion

Large distributed cortical systems are activated during the late cognitive evoked potentials N2, P3 and SW. These systems are concerned with the orientation of attention and the contextual integration of cognitive events. While it is difficult to measure activity in individual brain areas using cognitive EPs, they provide an excellent means for monitoring on-line the activation level in these core cortical systems for cognition.

Acknowledgements

We thank Drs. P. Baudena, G. Heit, M. Smith, J. Clarke and J. Stapleton for collaboration in the studies reported here. This work was supported by NIH (NS18741), INSERM, HFSPO, VA and ONR.

References (69)

  • M.E. Smith et al.

    Human medial temporal lobe potentials evoked in memory and language tasks

    Electroenceph. clin. Neurophysiol.

    (1986)
  • E. Snyder et al.

    Long-latency evoked potentials to irrelevant, deviant stimuli

    Behav. Biol.

    (1976)
  • N.K. Squires et al.

    Two varieties of long-latency positive waves evoked by unpredictable auditory stimuli in man

    Electroenceph. clin. Neurophysiol.

    (1975)
  • J.M. Stapleton et al.

    Endogenous potentials evoked in simple cognitive tasks: Depth components and task correlates

    Electroenceph. clin. Neurophysiol.

    (1987)
  • C. Alain et al.

    Human intracerebral potentials associated with target, novel and omitted auditory stimuli

    Brain. Topogr.

    (1989)
  • T. Allison et al.

    Face recognition in human extrastriate cortex

    J. Neurophysiol.

    (1994)
  • I. Altafullah et al.

    Focal medial temporal lobe spike-wave complexes evoked by a memory task

    Epilepsia

    (1988)
  • M.E. Bitterman

    The comparative analysis of learning

    Science

    (1975)
  • Chauvel, P., Vignal, J.P., Biraben, A., Badier, J.M. and Scarabin, J.M. Stereo-electroencephalography. In: G. Pawlick...
  • J.M. Clarke et al.

    Intracerebral measures of lateralized processing in humans: Effects of visual field, response hand, and processing stage

    Soc. Neurosci. Abstr.

    (1992)
  • J.M. Clarke et al.

    Auditory and visual sensory representations in human prefrontal cortex as revealed by stimulus-evoked spike-wave complexes

    Brain

    (1995)
  • D. Cohen et al.

    MEG versus EEG localization test using implanted sources in the human brain

    Ann. Neurol.

    (1990)
  • B.N. Cuffin et al.

    Tests of EEG localization accuracy using implanted sources in the human brain

    Ann. Neurol.

    (1991)
  • A.M. Dale et al.

    Improved localization of cortical activity by combining EEG and MEG with MRI cortical surface reconstruction: a linear approach

    J. Cogn. Neurosci.

    (1993)
  • A.M. Dale et al.

    Spatiotemporal imaging of coherent motion selective areas in human cortex

    Soc. Neurosci. Abstr.

    (1995)
  • Desmedt, J.E. Scalp-recorded cerebral event-related potentials in man as point of entry into the analysis of cognitive...
  • Donchin, E., McCarthy, G., Kutas, M. and Ritter, W. Event-related potentials in the study of consciousness. In: R.J....
  • Groll-Knapp, E., Ganglberger, J.A., Haider, M. and Schmid, H. Stereoelectroencephalographic studies on event-related...
  • Halgren, E. Evoked potentials. In: A.A. Boulton, G. Baker and C. Vanderwolf (Eds.), Neuromethods, Vol. 15:...
  • E. Halgren

    Firing of human hippocampal units in relation to voluntary movements

    Hippocampus

    (1991)
  • Halgren, E. and Marinkovic, K. Neurophysiological networks integrating human emotions. In: M. Gazzaniga (Ed.), The...
  • Halgren, E. and Marinkovic, K. General principles for the physiology of cognition as suggested by intracranial ERPs....
  • E. Halgren et al.

    Endogenous potentials generated in the human hippocampal formation and amygdala by infrequent events

    Science

    (1980)
  • Halgren, E., Stapleton, J.M., Smith, M.E. and Altafullah, I. Generators of the human scalp P3s. In: R.Q. Cracco and I....
  • Cited by (0)

    View full text